Datasets:

eswardivi

/

books

like 0

eam (in case of analog receive beamforming) during data and control reception [11]. The beams corresponding to the large antenna arrays are narrower and beam tracking may fail; therefore, beam recovery procedures have been specified in NR where a device can trigger a beam recovery procedure. Furthermore, a cell may consist of multiple transmission points, each transmitting its own beams; in that case, beam management procedures would allow device-transparent mobility for seamless handover between the beams of different transmission points. In addition, uplink-centric and reciprocity-based beam management is possible by utilizing uplink reference signals. The possibility of spatially separating users increases when using a large number of antenna elements in lower frequency bands in both uplink and downlink; however, this requires the knowledge of channel at the transmitter. In NR, extended support for such multi-user spatial multiplexing was introduced, either through high-resolution CSI feedback with a linear combination of DFT vectors, or uplink SRS improvements based on channel reciprocity. Moreover, support for distributed MIMO has been introduced, where the device can receive multiple PDCCHs and PDSCHs per slot to enable simultaneous data transmission from multiple transmission points to the same user [69]. The support of hybrid beamforming and high-resolution CSI feedback have been two impor- tant design principles in NR MIMO. The first goal is addressed by beam management, where a UE measures a set of analog beams for each digital port and reports the beam qual- ity. The gNB then assigns one or a small number of analog beams to the UE. As the down- link channel experienced by the UE changes, the gNB can modify this assignment, particularly when the link associated with the assigned beam fails. While beam management is instrumental in above 6 GHz frequency bands, it is also applicable to sub-6 GHz bands. For instance, in multipoint transmission scenario, where multiple transmission-reception points are asso

ciated with a UE, each link corresponds to a beam. The second goal is addressed by designing a modular and scalable CSI framework. High-resolution spatial channel information is provided via two-stage high-resolution precoding. The first stage involves the choice of basis subset, and the second stage comprises a set of coefficients for approximating a channel eigenvector with a linear combination of the basis subset. Note that while beam management and CSI acquisition can be operated independently, they can be used together for mobile UEs. In NR, the DM-RS and the TRS are used to estimate the timing offset and frequency errors. To guarantee that the timing-offset and frequency-error of the antenna panels can be esti- mated independently, the DM-RS and TRS are grouped, and the grouping information is signaled to the UE. The UE distinguishes the time-frequency resource of reference signals 588 Chapter 4 allocated for different data streams. In this way, timing-offset and frequency-errors can be independently estimated for different MIMO layers in non-coherent MIMO transmission to improve the performance. For downlink transmission with multiple antenna panels, the accuracy of CSI acquisition is critical for the system performance. To characterize the channel directional information, the codebook and the related feedback mechanism are typically designed for CSI acquisition, especially for FDD systems. For users in the cell edge of TDD systems, reciprocal channel estimation based on uplink reference signals is also inaccurate, and codebook-based CSI feedback can help improve the performance. Codebook design for massive MIMO should be flexible for different antenna array structures and applicable to various numbers of anten- nas. For a uniform panel array, the codebook is designed similar to the one used for FD- MIMO. While a single DFT vector can only characterize one spatial channel path, an advanced codebook taking the combination of two or more DFT vectors as the precoder can capture the characteristics of multipat

h channels. For non-uniform panel array, the antenna elements in the horizontal/vertical direction cannot be viewed as a uniform array. Thus array response vector is not in DFT form, and additional phase difference exists between the panels. In other words, the DFT vector cannot capture the actual channel response but can cause beam distortion and beam gain reduction. Moreover, panels at one transmit/ receive point are not easily calibrated in a practical implementation, where a fixed or ran- dom phase may exist among different panels. A good codebook design should use phase/ amplitude factors among panels to combine the beamforming vectors of the panels to match the array response [59]. In multi-panel MIMO codebook design, 2D-DFT vectors may be used as the per-panel beamforming vectors representing the spatial propagation properties for each panel. Furthermore, an inter-panel co-phasing factor is added across the DFT-based codeword to reduce performance loss. Let us use the two-panel case as an example, one can design the multi-panel codebook W for dual-polarized antenna array as W = [W|W1 W2W24] where W1 and W1 consist of DFT vectors reflecting long-term channel characteristics of antenna panel 1 and panel 2, respectively; W2 and W2 consist of phase factors reflecting the short-term and frequency-selective channel elements of panels 1 and 2, respectively, and P is the inter-panel co-phasing factor, which is designed for feedback in a wideband manner or subband manner according to the use cases. The precoding vectors for both panels are restricted to be the same in order to reduce the feedback overhead [59]. For a uniform or non-uniform planer array with ideal synchronization or small phase offset, coherent transmission can be used to achieve spatial multiplexing/diversity gain. However, the performance of coherent transmission may be degraded in practice due to different reflection and refraction propagation paths which cause different average channel delay and different average channel gain for a given directi

on of arrival and departure. If antenna ports from different panels experience different large-scale fading, it causes the eigenvectors of New Radio Access Physical Layer Aspects (Part 2) 589 the channel matrix to become ill-conditioned (rank deficiency), resulting in inaccurate CSI for coherent MIMO transmission. Moreover, time-varying amplitude or phase calibration error among the panels may occur in practice, especially when the antenna panels have independent clocks and different operating temperatures. In addition, the frequency offset for different panels observed at the receiver may be different when there is a time-varying relative phase between different antenna panels. The reference signals transmitted from geographically separated panels may experience dif- ferent Doppler shift and Doppler spread, because the UE may move in a different directions relative to the panels. If the UE assumes the panels have the same Doppler shift and Doppler spread, the frequency offset estimation would be inaccurate, which results in per- formance degradation. The relative frequency offset between panels can be from 0 to 300 Hz, due to the following factors [59]: Doppler shift: For a UE moving with the speed of 30 km/h at the carrier frequency of 2 GHz, the maximum difference between the experienced Doppler shifts is 111 Hz. Frequency error: In LTE, the maximum tolerable oscillator inaccuracy is +1 0.05 ppm for the wide area base station classes; thus the maximum frequency error between two non-calibrated panels with independent oscillators at the carrier frequency of 2 GHz is about 200 Hz. Thus the total amount of relative frequency offset is around 300 Hz. As the carrier frequency increases, the frequency offset problem would be more severe. The same timing offset assumption for different panels may impact the accuracy of chan- nel estimation. With a timing offset, a random linear phase factor will be added to the channel coefficient on adjacent subcarriers, which is difficult to accurately estimate. The inter-symbol in

terference caused by the timing offset impacts the channel estimation accu- racy in both CSI measurement and data demodulation, resulting in significant MIMO per- formance degradation. It can be shown that the performance loss due to timing offset is not negligible, especially in the case of negative timing offset. For positive timing offset, when the aggregated signals are received with the timing difference shorter than the cyclic prefix length, less ISI is incurred. For the negative timing offset, since the incurred ISI cannot be mitigated by removing the cyclic prefix, severe performance loss is caused. The capability of interference suppression in the receiver is critical for the performance of non-coherent MIMO transmission. For coherent MIMO transmission, a traditional linear receiver such as MRC/MMSE can be used to achieve acceptable performance, while for non-coherent MIMO transmission, an advanced receiver is required. If the data streams from different directions (especially in the case of widely spaced panels) arrive at the UE in different angles, a simple linear MMSE receiver may be sufficient. However, the per- formance gain depends on the remaining interference after the space-domain interference rejection by the MMSE processing. If different streams arrive at the UE from the same or adjacent directions (e.g., different antenna panels located in a centralized manner), Chapter 4 Downlink MIMO operation in sub-6 GHz Single CSI-RS Multiple CSI-RS SRS-based CSI-RS may be beamformed Combines beam selection with codebook Allows codebook feedback feedback (multiple beamformed CSI-RS with CRI Exploits TDD reciprocity feature Similar to LTE class-A CSI feedback (gNB feedback) Similar to SRS-based operation in LTE transmit CSI-RS; UE computes RI/PMI/CQI) Similar to LTE Class-B CSI Feedback (gNB Supports arrays with arbitrary number of TXRUs Maximum of 32 ports in the CSI-RS (codebooks transmits one or more CSI-RS, each in different are defined for up to 32 ports) directions; UE computes CRI/PMI/CQI) Procedur

e (UE transmits SRS, gNB computes precoding weights) Typically intended for arrays having 32 TXRUs or Supports arrays with arbitrary number of TXRUs less with no beam selection (no CRI) Maximum of 32 ports per CSI-RS CSI-RS (8 ports) CSI-RS (32 Ports) CSI-RS (8 ports) CSI-RS (8 ports) CSI-RS (8 ports) RI/PMI (32)/CQI RICQI CRI/PMI/CQI Downlink MIMO operation in above 6 GHz Single-panel array Multipanel array Combination of analog beamforming and digital precoding at baseband Combination of analog beamforming and digital precoding at baseband Analog beamforming is typically one RF beamforming weight-vector per Analog beamforming is typically one RF beamforming weight-vector per polarization per panel polarization (a single cross-polarized beam) One cross-polarized beam per subpanel Supports two TXRUs and single-user MIMO only Number of TXRUs = 2 X number of panels Digital precoding options: None (rank-2 all the time); CSI-RS based (RI/PMI/ Digital precoding options: CSI-RS based (RI/PMI/CQI); SRS-based (RI/CQI); CQI); and SRS-based (RI/CQI) and SU-MIMO and MU-MIMO (typically one UE per cross-polarized beam) SU-MIMO SU-MIMO MU-MIMO TXRUs at the gNB Up to 4 UEs at a time 2 ports/UE 1< rank <2 2 TXRUs at the gNB One UE at time 1< rank< 2 8 TXRUs at the gNB One UE at time 8 ports/UE 1< rank< 8 Figure 4.71 Summary of NR downlink multi-antenna operation [26]. inter-stream interference would be inevitable, and it would be the dominant factor to degrade the demodulation performance. In this case, a nonlinear receiver may be needed. When the received signal power difference among the streams is larger than 3 dB, the codeword-level successive interference cancellation or SIC receiver can satisfactorily perform. However, when the received signal power is almost the same for different data streams, the performance of SIC will deteriorate. In that case, a parallel interference can- cellation receiver would be a more suitable choice. In NR, the improved DM-RSs and tracking reference signals can be used to estimate the timing of

fset and frequency errors in above 6 GHz frequencies [59]. The NR downlink multi-antenna operation has been dis- cussed in different sections of this chapter and is summarized in Fig. 4.71. New Radio Access Physical Layer Aspects (Part 2) 591 UE MAC gNB MAC UL-SCH/UCI [PUSCH] UCI [PUCCH] Transport Blocks CRC Calculation + Attachment Sequence Generation UCI Bit Sequence Generation and Scaling LDPC Graph Selection UCI Channel Encoding Code Block Segmentation Configuration HARQ ACK/NACK Channel Coding (LDPC Encoding) Modulation and PUCCH format- Channel Decoding (LDPC Rate Matching + HARQ Specific Processing decoding) + HARQ Buffer HARQ ACK/NACK Data and Control Multiplexing PUCCH De-scrambling + de-interleaving Configuration Scrambling and Modulation Demodulation Layer Mapping Layer De-Mapping CSI Reports/ Measurements Transform Precoding (DFT) IDFT (Transform Precoding) CSI Estimation Precoding MIMO Detection/Equalization Virtual/Physical resource Channel Mapping Resource De-mapping Estimation OFDM Modulation + CP Reference Signal Generation Reference Signal CP Removal + OFDM PUCCH Format- Insertion DM-RS/SRS/PT-RS Configuration Demodulation Specific Decoding TX Beamforming RX Beamforming IQ Modulation/RF Up-conversion/ LNA and RX Filter/IQ PA and TX Filter Demodulator/Down-conversion Figure 4.72 Overall uplink physical layer processing 5]. 4.2 Uplink Physical Layer Functions and Procedures 4.2.1 Overall Description of Uplink Physical Layer The NR uplink physical-layer consists of higher layer configurable functional blocks that are configured according to the uplink physical channel characteristics, use case, deploy- ment scenario, etc. As shown in Fig. 4.72, the physical-layer processing generally includes receiving higher layer data (e.g., MAC PDUs in the case of uplink shared channel); CRC calculation and attachment; channel encoding and rate matching; modulation; mapping to physical resources and antennas; multi-antenna processing; and support of layer-1 control and HARQ-related signaling. The physical-layer

model for RACH transmission is character- ized by a physical RACH (PRACH) preamble format that consists of a cyclic prefix, a pre- amble, and a guard time during which no signal is transmitted. 592 Chapter 4 4.2.2 Reference Signals 4.2.2.1 Demodulation Reference Signals Uplink DM-RSs are used for channel estimation and, as shown in Fig. 4.73, they are subject to the same precoding as uplink shared channel. Uplink reference signals are required to have small power variation in the frequency domain to allow similar channel-estimation quality for all frequencies spanned by the reference signals. This requirement is fulfilled for CP-OFDM waveform by using a pseudo-random sequence with good autocorrelation prop- erties. However, for DFT-precoded OFDM waveform, limited power variations as a func- tion of time are also important to achieve signal transmission with a low cubic metric. Furthermore, sufficient number of reference signal sequences of a given length, correspond- ing to a certain reference signal bandwidth, should be available in order to avoid restrictions when scheduling multiple devices in different cells. It is shown that Zadoff-Chu sequences can satisfy these requirements. From a Zadoff-Chu sequence with a given group index and sequence index, additional reference signal sequences can be generated by applying differ- ent linear phase rotations in the frequency domain. The DFT-precoded OFDM waveform in the uplink only supports single-layer transmission and is primarily specified for limited link-budget scenarios. Due to the importance of low cubic metric property and the corresponding high-power-amplifier efficiency, the reference signal structure is somewhat different compared to the CP-OFDM uplink case. In general, transmitting reference signals that are frequency-multiplexed with other uplink transmis- sions from the same device is not suitable for the uplink as it would negatively impact the device power-amplifier efficiency due to increased cubic metric. Instead, certain OFDM Modulated symbols Layer

mapping DM-RS Precoding W Mapping to DM-RS antenna ports Resource mapping Mapping to SRS antenna ports Antenna mapping (Logical to physical antenna mapping) Figure 4.73 Processing of uplink DM-RS with PUSCH. New Radio Access Physical Layer Aspects (Part 2) 593 symbols within a slot are used exclusively for DM-RS transmission, that is, the reference signals are time-multiplexed with the data transmitted on PUSCH from the same device. The NR uses the same DM-RS structure for downlink and uplink in the case of CP-OFDM waveform. For DFT-spread OFDM waveform in the uplink, the DM-RS is based on Zadoff-Chu sequences and supports contiguous allocations and single-layer transmission, similar to LTE, in order to improve the power-amplifier efficiency. Multiple orthogonal ref- erence signals can be generated in each DM-RS occasion where different reference signals are separated in the time, frequency, and code domains. Two different types of DM-RSs can be configured, namely, DM-RS Type 1 and Type 2, which differ in the maximum num- ber of orthogonal reference signals and the mapping to the resource elements in the frequency domain. Type 1 provides up to four orthogonal reference signals using a single- symbol DM-RS and up to eight orthogonal reference signals using a double-symbol DM- RS, whereas Type 2 provides 6 and 12 patterns for single and double-symbol DM-RS, respectively. The DM-RS type 1 or 2 should not be confused with the mapping type A or B (see Section 4.1.2.1), as different mapping types can be combined with different reference signal types. In the following sections, we describe the DM-RSs corresponding to PUSCH and various PUCCH formats for the CP-OFDM uplink waveform. 4.2.2.1.1 PUSCH DM-RS The DM-RS sequence for PUSCH rdd-rs(m) is generated according to = - 2c(2m)) + j(1 - 2c(2m + 1))]/2 where c(i) is a length-31 Gold sequence generated by the pseudo-random sequence generator defined in [6] and initialized with l is the OFDM symbol number within the slot, Nslot is the slot number within a frame, and are giv

en by the higher layer parameters scramblingID0 and scramblingIDI in the DMRS-UplinkConfig information element, respectively. The parame- ter NSCID E{0,1} may be signaled through the DM-RS initialization field of DCI format 0_1, otherwise, NSCID [6]. The sequence r'DM-RS(n) is mapped to an intermediate quantity bill according k = 4n + 2k' + A for DMR-RS Type 1 and k + k' + A for DMR-RS Type 2, and k' = 0, 1; l=1+1'; n E N; p = 0, ,NAP - 1 [6]. The spreading sequences wf(k'), wt(l)), and the offset A are given in Table 4.22 [6]. The intermediate quantity is precoded and multiplied with the amplitude scaling factor DM-RS QPUSCH in order to adjust the transmit power of the reference signals and are then mapped to the physical resources according to In the latter equation, W is the precoding matrix, 594 Chapter 4 Table 4.22: Parameters for PUSCH DM-RS Types 1 and 2 [6]. CDM Group wf(k') wt(I') k' = 0 k' = 1 DM-RS Type 1 DM-RS Type 2 it is assumed that the resource elements b(p,H are within the common resource blocks allo- cated for PUSCH transmission. It is further assumed that the reference point for the fre- quency index k is subcarrier 0 in common resource block 0; the reference point for time index l and the position lo of the first DM-RS symbol depends on the mapping type, that is, for PUSCH mapping type A, l is defined relative to the start of the slot, if frequency hopping is disabled; otherwise, it is measured relative to the start of each hop, and lo is given by the higher layer parameter For PUSCH mapping type B, l is defined rela- tive to the start of the scheduled PUSCH resources, if frequency hopping is disabled and rel- ative to the start of each hop otherwise and lo = 0 [6]. In other words, the pseudo-random sequence corresponding to DM-RS Type 1 is mapped to every second subcarrier in the frequency domain over the OFDM symbol assigned to DM-RS transmission. As shown in Table 4.22, antenna ports 0 and 1 are associated with CDM group 0 and mapped to even-numbered subcarriers and are separated in the cod

e- domain using different orthogonal sequences. The DM-RS corresponding to PUSCH for antenna ports 2 and 3 belong to CDM group 1 and are generated in the same way using odd-numbered subcarriers which are separated in the code domain. If more than four New Radio Access Physical Layer Aspects (Part 2) orthogonal antenna ports are needed, two consecutive OFDM symbols are used. The above DM-RS structure is used over each OFDM symbol and a length-2 orthogonal sequence is applied across time, resulting in up to eight orthogonal sequences. DM-RS Type 2 has a similar structure as Type 1; however, there are some differences, most notably the number of antenna ports supported. As shown in Fig. 4.74, each CDM group for DM-RS Type 2 consists of two neighboring subcarriers over which a length-2 orthogonal sequence is applied to separate the two antenna ports sharing the same set of subcarriers. Two such pairs of subcarriers are used in each resource block for one CDM group. Since there are 12 subcarriers in a resource block, up to three CDM groups each with two orthogonal reference signals can be created using one resource block over one Group-0 >AP-0 Group-0 Group-1 Group-1 Group-2 Single-symbol Type 2 DM-RS Single-symbol Type 1 DM-RS Group-0 Group-0 Group-1 Group-1 Group-2 Group-0 Group-0 Group-1 Group-1 AP-10 Group-2 AP-11 Double-symbol Type 2 DM-RS Double-symbol Type 1 DM-RS Figure 4.74 Illustration of DM-RS Type-1 and Type-2 time-frequency-code structures [14]. 596 Chapter 4 OFDM symbol. The number of Type 2 DM-RSs can be increased up to 12 by applying a length-2 orthogonal sequence across time domain. While the basic structures of DM-RS Type 1 and Type 2 are similar, the frequency-domain density of DM-RS Type 1 is higher than that of Type 2; however, Type 2 provides a larger number of orthogonal patterns, which useful for MU-MIMO use cases. The type of the reference signal structure is deter- mined by dynamic scheduling and higher layer configuration. 4.2.2.1.2 PUCCH DM-RS The DM-RS sequence for PUCCH format 1 is define

d = wi(m)r(a)(n) where n = 0, 1, NRB - 1, m = 0, 1, and m' =0 when intra - slot frequency hopping is enabled; otherwise, m' In the latter equation, the number of the DM-RS symbols NSF,m', orthogonal sequence wi(m), and the pseudo-random sequence (n) are given in [6]. The DM-RS sequence is multiplied by the scaling factor BPUCCH,1 in order to adjust the transmit power and is sequentially mapped starting with z(0) to resource elements (k, 1) in a slot on a single-antenna port such that = l = 0, 2, 4, wherein l = 0 corresponds to the first OFDM symbol of the PUCCH transmis- sion and (k,1) is within the resource blocks assigned for PUCCH transmission [8]. The DM-RS sequence corresponding to PUCCH format 2, IDD-RS(m, I) is generated as - sequence c(i) is mod 231. Index l is the OFDM symbol number within the slot, nslot is the slot number within the radio frame, and NO {0, 65535}. The PUCCH format 2 DM-RS sequence scaled with the scaling factor BPUCCH,2 to adjust the transmit power [8] and is sequentially mapped starting with r(0) to resource elements (k,1) in a slot on a single-antenna port such that akill = k = 3m + 1 where the frequency index k is defined relative to subcarrier 0 of common resource block 0 and (k,l) is within the resource blocks assigned for PUCCH transmission [6]. The DM-RS sequence corresponding to PUCCH format 3/4 r'DM-RS(m,I) is generated according to r'DM-RS(m, = 0, 1, ,MPUCCH,s 1 where the number of subcar- riers for this PUCCH format MPUCCH,S and the pseudo-random are given in [6]. The cyclic shift a varies with the symbol number and slot number with mo = 0 for PUCCH for- mat 3. The PUCCH format 3/4 DM-RS sequence is multiplied by the scaling factor BPUCCH,S wherein S E {3,4} to adjust the transmit power and is sequentially mapped starting with to resource elements (k,l) on a single-antenna port such that m=0,1,. MPUCCH,s - 1. The frequency index k is defined relative to subcarrier 0 of the lowest numbered resource block assigned for PUCCH transmission, New Radio Access Physical Layer Aspects

(Part 2) 597 Table 4.23: DM-RS positions for PUCCH format 3 and 4 [6]. PUCCH Length DM-RS Position / within PUCCH Allocation No Additional DM-RS Additional DM-RS No Hopping Hopping No Hopping Hopping 1,3,6,8 1,3,6,9 1,4,7,10 1,4,7,11 1,5,8,12 and time index l is given in Table 4.23 with and without intraslot frequency hopping as well as with and without additional DM-RS, where l = 0 corresponds to the first OFDM symbol of the PUCCH transmission [6]. 4.2.2.2 Phase-Tracking Reference Signals There is typically a mismatch between the oscillator frequencies of the transmitter and receiver in a communication system, resulting in a shift of the received signal spectrum at the baseband. In OFDM, this effect creates a misalignment between the FFT bins and the peaks of the sinc(.) pulses of the received signal, which would compromise the orthogonality between the subcarriers and would result in a spectral leakage. Each subcarrier interferes with other subcarriers, although the effect is dominant between adjacent subcarriers. Since there are many subcarriers, this is a random process is equivalent to a noise with Gaussian distribu- tion. This random frequency offset degrades the SINR of the receiver. Therefore, an OFDM receiver will need to track and compensate the phase noise. The PT-RS is mainly introduced to compensate the CPE; however, PT-RS can also be used for ICI mitigation in higher fre- quency bands and potentially for CFO and Doppler estimation [54]. There is a trade-off between phase tracking accuracy and signaling overhead. If the density of PT-RS is high, phase tracking accuracy is high, and the CPE can be better compensated to achieve better performance. However, higher PT-RS density also means larger signaling over- head, which might lead to lower spectral efficiency or effective transmission rate. The studies show that the reduction of PT-RS density in time domain will degrade the BLER performance regardless of modulation order. However, the performance degradation is particularly signifi- cant when time d

ensity is reduced from 1 to 2, that is, from PT-RS for each OFDM symbol 598 Chapter 4 to PT-RS for every other OFDM symbol in the case of 256 QAM [54]. In that case, although signaling overhead is halved for time density 2, that is, more information bits can be transmit- ted in each RB, the effective data transmission rate suffers performance loss due to degraded BLER. In contrast, for 64QAM the degradation due to time density reduction is much less. Therefore, the time density of PT-RS can be a function of modulation order and it should increase with higher modulation order. Since the receiver only needs to track the phase difference using the PT-RS, it does not need to know the amplitude of the PT-RS. Therefore, unlike other reference signals, for example, DM-RS, where the reference signals are formed by a pseudo-random sequence of symbols, PT-RS reference signals can use the exact same symbol. In MU-MIMO, PT-RS can be configured for each user and it is possible that the same subcarrier is used for multi- ple users, thus PT-RS collisions may happen, which would degrade the CPE compensation for two reasons: (1) the interference pattern is not completely random since the same sym- bol is used for PT-RS and (2) the interference level is higher when the power of PT-RS is boosted for more accurate CPE compensation. In such cases, it would be better to avoid PT-RS collision SO that the interference is randomized without power boosting. This can be avoided by introducing an RB-level offset when configuring PT-RS for each user [54]. In NR, both DFT-S-OFDM and CP-OFDM waveforms are supported for the uplink transmis- sions and PT-RS signals are necessary for both waveforms. The PT-RS insertion follows a com- mon framework for both downlink and uplink in case of CP-OFDM waveforms. In the case of DFT-S-OFDM, two types of insertion mechanisms for PT-RSs were studied, that is, pre-DFT and post-DFT insertion. In the former mechanism, PT-RS signals are inserted in the frequency domain before DFT precoding SO that the resulting

waveform still maintains the single-carrier properties. In the latter mechanism, the PT-RS are inserted after DFT precoding of the data symbols via various mechanisms such as puncturing. The PAPR of such a mechanism can how- ever be controlled by using some signal processing techniques. In the time domain, the PT-RS locations can be configured to be either present in every symbol or every other symbol. The NR supports pre-DFT PT-RS insertion to conserve the single carrier property. In the frequency domain, PT-RSs are transmitted in every second or fourth resource block, resulting in a sparse frequency-domain structure. The density in the frequency domain corre- sponds to the scheduled transmission bandwidth such that the higher the bandwidth, the lower the PT-RS density. For the smallest bandwidths, no PT-RS is transmitted. To reduce the risk of collisions between PT-RSs associated with different devices scheduled on overlapping time- frequency resources, the subcarrier number and the resource blocks used for PT-RS transmis- sion are determined by the C-RNTI of the UE. The antenna port used for PT-RS transmission is given by the lowest numbered antenna port in the DM-RS antenna port group [14]. In a multi-TRP deployment, a UE can be supported by multiple co-located or non-co-located transmission points belonging to the same or different gNBs. The NR further supports large number of antenna elements at the gNB. These antenna elements are typically grouped as New Radio Access Physical Layer Aspects (Part 2) 599 panels, where the signals feeding the antenna panels are generated by separate oscillators, which require individual phase noise compensation. In designing PT-RS for multi-TRP deploy- ments, the orthogonal time-frequency allocation of PT-RS is crucial to minimize interference. The orthogonality of the PT-RS can be ensured in the frequency domain via frequency-division multiplexing of the PT-RS bearing resource elements. The increase in signaling overhead is a valid concern when there is a higher number of g

NBs or many transmit panels. The typical CPE caused by the phase noise rotates the constellations by a limited margin, SO only the higher order modulation schemes are impacted by the CPE. The users with higher MCS receive good SNR levels and are usually located closer to the gNB. When the MU-MIMO users are grouped, there will be higher and lower MCS users in these groups. If the PT-RS is transmitted without power boosting and with a wider beam than the narrow-beam data transmissions, the received EIRP for the PT-RS will be lower relative to the data transmissions. The lower MCS users that are generally further away from the cell center will receive PT-RS with a much lower effective power and will be able to discard PT-RS as interference, that is, they will not need CPE correc- tion. They will be able to request the gNB to allocate data within these resource elements trans- mitted through narrow-beams. The same resource elements are used for the PT-RS in the wider beam transmissions for the benefit of higher MCS users, for whom the same resource elements will not be utilized in the narrow-beam data transmissions. With this effective power discrimi- nation, non-orthogonal multiplexing of PT-RS and data is possible for the MU-MIMO config- urations, which effectively increases the system spectral efficiency [54]. In CP-OFDM uplink, the precoded phase-tracking reference signal for subcarrier k on layer j is given by r'pT-Rs(p,m) =[(1-2c(2m)) = +j(1-2 1))]/2 if p=p or p=p" where antenna ports p' or (p',p") are associated with PT-RS transmission. The pseudo-random sequence initialized 2Nnscid + nscid] mod 231, in which l is the OFDM symbol number within the slot, nslot is the slot number within a frame, and NID, N1D E {0, 1, 65535} are given by the higher layer para- meters scramblingID0 and scramblingID1, respectively. The parameter NSCID E {0, may be signaled through the DM-RS initialization field of DCI format 0_1; otherwise, NSCID = 0 [6]. The UE transmits PT-RSs (if configured) only in the resource blocks designate

d to PUSCH transmission. The PT-RS is mapped to resource elements according BPT-RsW[rPT-Rs(po,2n+k') TPT-RS(Pv-1,2n+k)) where k + 2k' for configura- tion Type 1 or k=6n+k'+ A for configuration Type 2, if l is within the OFDM symbols allocated for the PUSCH transmission, the resource element (k,l) is not used for DM-RS. parameters as well as the precoding matrix W are given in [6]. The configuration type is provided by the higher layer parameter DMRS-UplinkConfig. In the preceding expression, scaling factor BPT-RS is used to adjust the transmit power. The set of time indices l is defined relative to the start of the PUSCH allocation is defined by max(lref+(i- 1)LPT-RS + 1, lref) lref + iLPT-RS ALPT-RS 6 {1,2,4} where any symbol in this interval which overlaps with a DM-RS symbol is skipped. 600 Chapter 4 Table 4.24: Time-domain/frequency-domain density of phase tracking reference signal as a function of scheduled modulation coding scheme/ bandwidth [9]. Scheduled MCS Time-Domain Density LPT-RS 0 < MCS < MCS1 No PT-RS MCS < MCS < MCS2 Every OFDM symbol MCS2 < MCS < MCS3 Every second OFDM symbol MCS3 < MCS < MCS4 Every fourth OFDM symbol Scheduled Bandwidth Frequency Domain Density KPT-RS <NRB NRB1 No PT-RS NRB1 VI NRB <NRB Every second RB NRB2 VI NRB Every fourth RB PUSCH_(data/ PUSCH_(data/ PUSCH_(data/ control control control) PT-RS on every OFDM symbol PT-RS on every second OFDM symbol PT-RS on every fourth OFDM symbol PT-RS on every second resource block PT-RS on every fourth resource block Scheduled bandwidth Figure 4.75 Example CP-OFDM uplink time-frequency resource mappings. The resource blocks allocated for PUSCH transmission are numbered from 0 to NRB-1 rela- tive to the lowest scheduled resource block for the purpose of PT-RS transmission. The subcarriers associated with these resource blocks are numbered in increasing order from 0 to NRBNRB starting with the lowest frequency. The PT-RS is mapped to subcarriers k=kref where KPT-RS &{2,4} (see Table 4.24). The parameter KRB=NRNTI mod KPT-RS, if NRB mod KPT

-RS 0; otherwise, kref NRNTI mod(NRB mod KPT-RS). The parameter NRNTI is the RNTI associated with the DCI that is used to schedule PUSCH [6]. Example time-frequency mapping of CP-OFDM uplink PT-RS is depicted in Fig. 4.75. New Radio Access Physical Layer Aspects (Part 2) 601 4.2.2.3 Sounding Reference Signal A UE can be configured to transmit SRS in order to enable the gNB to estimate the uplink channel. Similar to the downlink CSI-RS, the SRS can serve as QCL reference for other physi- cal channels such that they can be configured and transmitted quasi-co-located with SRS. As a result, if the knowledge of a suitable receive beam for the SRS is available, the receiver would know that the same receive beam should be suitable for other physical channels. The SRS sup- ports up to four antenna ports, and it is designed to have low cubic metric, enabling efficient operation of the high-power amplifier. In general, the SRS can span one, two, or four consecu- tive OFDM symbols and is located within the last six symbols of a slot. In the frequency domain, an SRS occasion has a comb structure where the SRS is transmitted on every Nth sub- carrier where N = 2 or 4, referred to as comb-2 or comb-4. The SRS transmissions from dif- ferent devices can be frequency-multiplexed within the same frequency range using different comb patterns corresponding to different frequency offsets. For comb-2, that is, transmitting SRS on every other subcarrier, two SRSs can be frequency multiplexed, whereas for comb-4, up to four SRSs can be frequency multiplexed. Fig. 4.76 illustrates example SRS multiplexing assuming a comb-2 structure spanning two OFDM symbols. The sequences used to represent a set of SRS are based on Zadoff-Chu sequences. A Zadoff-Chu sequence is a complex-valued sequence with constant amplitude property whose cyclically shifted versions exhibit low cross-correlations. Thus under certain conditions, the cyclically shifted versions of each sequence remain orthogonal to one another. A Zadoff-Chu sequence that has not been s

hifted is referred to as a root sequence. The uth root Zadoff-Chu sequence of prime length N is defined as follows: 0<n<N- 1 (N is an odd integer) 1 (N is an even integer) where N is an integer, denoting the length of the Zadoff-Chu sequence. One can verify that u(n) is periodic with period N, that is, xu(n) = xu(n+N), In. In other words, the sequence index uis a prime relative to N. For a fixed value of the Zadoff-Chu sequence has an ideal periodic autocorrelation property (i.e., the peri- odic autocorrelation is zero for all time shifts other than zero). For different values of index the Zadoff-Chu sequences are not orthogonal, rather exhibit low cross-correlation. If the sequence length N is selected as a prime number, there are different sequences with periodic cross-correlation of /VN between any two sequences regardless of time shift. The Zadoff-Chu sequences are a subset of constant amplitude zero autocorrelation sequences. The properties of Zadoff-Chu sequences can be summarized as follows: They are periodic with period N, if N is a prime number, that is, xu(n + N) = xu(n). Given N is a prime number, the DFT of a Zadoff-Chu sequence is another Zadoff-Chu sequence con- jugated and time-scaled multiplied by a constant factor, that is, Xu[k] = where V is the multi- plicative inverse of modulo N. It can be shown that x"(vk) = x",(k)ein(1-v)k/N. The autocorrelation of a prime-length Zadoff-Chu sequence with a cyclically shifted version of itself also yields zero autocorrelation sequence, that is, it is non-zero only at one instant which corresponds to the cyclic shift zero. The cross-correlation between two prime-length Zadoff-Chu sequences, that is, different , is constant and equal to 1/VN The Zadoff-Chu sequences have low PAPR. 602 Chapter 4 PUSCH (data (data Active control PUSCH bandwidth part control SRS Comb-2 structure for two UEs General SRS/PUSCH multiplexing Figure 4.76 Example illustration of NR SRS structure in time and frequency domains [6]. Although Zadoff-Chu sequences of prime length are prefer

red in order to maximize the number of available sequences, the SRS sequences are not of prime length. The SRS sequences are extended Zadoff-Chu sequences based on the longest prime-length Zadoff-Chu sequence with a length N less than or equal to the desired SRS sequence length. The sequence is then cyclically extended in the frequency domain up to the desired SRS sequence length. As the extension is performed in the frequency domain, the extended sequence has a constant spectrum and a perfect cyclic autocorrelation, but the time-domain amplitude is not constant and slightly varies. The extended Zadoff-Chu sequences are used as SRS sequences for sequence lengths of 36 or larger, corresponding to an SRS extending over 6 and 12 resource blocks in the cases of comb-2 and comb-4, respectively. For shorter sequence lengths, special flat-spectrum sequences with good time-domain envelope proper- ties are found through computer search since there would not be sufficient number of Zadoff-Chu sequences available. An SRS resource is configured by the SRS-Resource information element and consists of 1, 2, or 4 antenna ports, where the number of antenna ports is set by the higher layer parameter nrofSRS-Ports, and NSRS symb E {1,2,4} consecutive OFDM symbols provided via parameter nrofSymbols contained in the higher layer parameter resourceMapping. The starting position lo in the time domain given by lo = Nsymb - 1 - offset where the offset parameter 5} counts symbols backwards from the end of the slot and is given by the field startPosition contained in the higher layer parameter resourceMapping. The SRS sequence for an SRS resource is generated as NSRS where the length of the sequence is given by MRSmssc(a(n) sc,b is a Zadoff-Chu sequence, 8 = log2(KTC) in which KTC denotes the transmission comb number given by the higher layer parameter transmissionComb. The cyclic shift Ai for antenna port Pi is given New Radio Access Physical Layer Aspects (Part 2) 603 where Na100oa - is provided by the higher layer parameter transmissio

nComa maximum number of cyclic shifts is NSRS max = 12 when KTC = 4 and NSRS max =8 when KTC=2_[6]. The sequence group is defined as = and the sequence number V depends on the higher layer parameter groupOrSequenceHopping in the SRS-Config information element. Furthermore, the SRS sequence identity nSRS is given by the higher layer parameter sequenceId in the SRS-Config information element. If groupOrSequenceHopping parameter indicates neither group nor sequence hopping, fgh(nslot, l') = and v=0 are used; otherwise, if parameter groupOrSequenceHopping indicates groupHopping, group hopping is used and fgh(nslot,l' = 30 and V = 0 where the pseudo-random sequence c(i) is initialized with Cinit nSBS at the beginning of each radio frame. If parameter groupOrSequenceHopping indicates sequenceHopping, sequence hopping is used and fgh(nslot, l') = 0 and V = for ISRS - 6NRB , otherwise, v=0. The pseudo-random sequence c(i) is initialized similar to the group hopping case [6]. Each SRS is transmitted on a designated SRS resource, and the SRS sequence r(pi)(n, l') for each OFDM symbol l' and antenna port pi is multiplied by a scaling factor BSRS to adjust its transmit power. The scaled sequence is then sequentially mapped to resource elements (k,1) starting with r(pi)(0, 1') in a slot for each antenna port such that = 0,1, , MRS and = 0,1, NSRS - 1 The length of the SRS sequence is [6]. The frequency-domain starting position k(p) defined bykicM where BSRS denotes the SRS + KTC/2) mod KTC if max /2,..., NSRS max NSRS otherwise, KLPI = ktc [6]. The frequency domain shift value nshift adjusts the SRS allocation with respect to the common resource block grid. The transmission comb offset KTC E {0, 1, KTC - 1} is con- tained in the higher layer parameter transmissionComb in the SRS-Config information element and nb is a frequency position index. Frequency hopping of the SRS is configured by the parameter bhop E {0, 1, 2, 3}. If bhop > BSRS, frequency hopping is disabled and the frequency position index nb is set to a constant as

follows nb = AnRRC/MSRS,b mod Nb for all NSRS symb OFDM symbols of the SRS resource. The value of the parameter NRRC is given by the higher layer parameter freqDomainPosition and the values of MSRS,b and Nb for b = BSRS are given in [6]. If bhop < BSRS, the frequency hopping is enabled and the frequency position indices nb are defined as nb = 4nRRC/MSRS,b mod Nb if b < bhop; otherwise, nb = {Fb(nsRs) + Nb where Nb and Fb(nsRs) are defined in [6]. The quantity NSRS counts the 604 Chapter 4 number of SRS transmissions. For the case of an SRS resource configured as aperiodic by the higher layer parameter resourceType, NSRS I//R within the slot where NSRS symb symbol SRS resource is transmitted. The quantity a repetition factor configured via higher layer signaling. For the case of an SRS resource configured as periodic or semi-persistent by the higher layer parameter resourceType, the value of the SRS counter is given by slots satisfy Notamen frame TSRS =0, where TSRS is SRS periodicity defined in the number of slots and Toffset is the slot offset [6]. Note that when supporting more than one SRS antenna port, different antenna ports share the same set of resource elements and the same baseline SRS sequence; nevertheless, different phase rotations are applied to separate them. Applying a phase rotation in the frequency domain is equivalent to cyclic shift in the time domain. As we mentioned earlier, the SRS can be configured as periodic, semi-persistent, or aperi- odic transmission A periodic SRS is transmitted with a certain configured periodicity and a certain configured slot offset within that period. A semi-persistent SRS has a configured periodicity and slot offset in the same way as a periodic SRS; however, the SRS transmis- sion is performed according to the configured periodicity and slot offset that is activated or deactivated via MAC control element signaling. An aperiodic SRS is only transmitted when explicitly triggered by means of a DCI. It should be noted that SRS activation/deactivation or triggering

for semi-persistent and aperiodic cases is not done for a specific SRS, rather for an SRS resource set which may include multiple SRSs [14]. A UE can be configured with one or several SRS resource sets, where each resource set includes one or more configured SRSs. All SRS occasions included within a configured SRS resource set are of the same type. In other words, periodic, semi-persistent, or aperiodic trans- mission is a property of an SRS resource set. A UE can be configured with multiple SRS resource sets that can be used for different purposes, including both downlink and uplink multi-antenna precoding and/or downlink and uplink beam management. The transmission of the set of configured SRS included in an aperiodic SRS resource set is triggered by a DCI. More specifically, DCI format 0_1 containing uplink grant and DCI format 1_1 containing downlink scheduling assignment include a 2-bit SRS-request that can trigger the transmission of one of the three different aperiodic SRS resource sets configured for the UE and the fourth bit combination corresponds to no trigger. The SRS antenna ports are typically not mapped directly to the UE's physical antennas, rather via some antenna mapping scheme. In order to provide connectivity regardless of the rotational direction of the device, the NR devices supporting high-frequency operation will New Radio Access Physical Layer Aspects (Part 2) 605 typically include multiple antenna panels pointing in different directions. The SRS may be mapped to one of those panels and transmission from different panels will then correspond to different antenna mapping schemes. The antenna mapping scheme has a real impact despite the fact that it is transparent to the gNB receiver; thus it is seen as an integral part of the overall channel from the UE to the gNB. The gNB may estimate the channel based on SRS transmission from a UE and subsequently select a precoding matrix that the device should use for uplink transmission. The device is then assumed to use that precoding matrix in combi

configuring one or more spatial correspondence between PUCCH and downlink physical Table 4.25: PUCCH formats [6]. PUCCH Number of Number Description Format OFDM Symbols of Bits PUCCH Short Short PUCCH of one or two symbols with small UCI PUCCH payloads of up to 2 bits with UE multiplexing in the same PRB Short PUCCH of one or two symbols with large UCI payloads of more than 2 bits with no multiplexing in the same PRB Long PUCCH of 4-14 symbols with small UCI payloads PUCCH of up to 2 bits with multiplexing in the same PRB Long PUCCH of -14 symbols with medium UCI payloads with some multiplexing capacity in the same Long PUCCH of 4-14 symbols with large UCI payloads with no multiplexing capacity over the same PRB 606 Chapter 4 signals such as CSI-RS or SS block. As a result, the device can transmit PUCCH using the same beam as it used for receiving the corresponding downlink signal. For example, if the spatial relation between PUCCH and SS block is configured, the device will transmit PUCCH using the same beam as it used for receiving the SS block. Multiple spatial relations can be configured and selected via MAC control elements [71]. It will be shown later in this section that the short PUCCH format of up to two UCI bits is based on sequence selection, while the short PUCCH format of more than two UCI bits fre- quency multiplexes UCI and DM-RS. The long PUCCH formats time-multiplex the UCI and DM-RS. Frequency hopping is supported for long PUCCH formats and for short PUCCH formats of duration two symbols. Long PUCCH formats can be repeated over mul- tiple slots. The UCI multiplexing in PUSCH is further supported when UCI and PUSCH transmissions coincide in the same slot, that is, UCI carrying HARQ-ACK feedback with 1 or 2 bits is multiplexed by puncturing and rate-matching PUSCH. In all other cases, the UCI is multiplexed by rate matching PUSCH. The UCI may carry CSI, HARQ ACK/ NACK, or scheduling request (SR). The QPSK modulation is used for long PUCCH with 2 or more bits of information, and short PUCCH with mo

re than 2 bits of information. The BPSK modulation is used for long PUCCH with a single information bit. Transform precod- ing is applied to long PUCCH [11]. In deployment scenarios where a SUL is used the UE is configured with two uplink carriers and one downlink carrier in the same cell. The transmissions on those uplink carriers are controlled by the network to avoid overlapping PUSCH/PUCCH transmissions in the time domain. The overlapping transmissions on PUSCH are avoided with properly scheduling the uplink transmissions, while overlapping transmissions on PUCCH are circumvented by proper configuration, that is, PUCCH can only be configured for one of the two uplink car- riers of the cell. In addition, initial access is supported on each of those uplink carriers [11]. A UE is semi-statically configured via RRC signaling to perform periodic CSI reporting using PUCCH and can be configured for multiple periodic CSI reports corresponding to one or more CSI reporting setting indications, where the associated CSI measurement links and CSI Resource Settings are also configurable. Periodic CSI reporting on PUCCH formats 2, 3, 4 supports Type I CSI with wideband granularity. A UE performs semi-persistent CSI reporting on PUCCH after successfully decoding a selection command, which contains one or more reporting setting indications where the associated CSI measurement links and CSI resource settings are configured. Semi-persistent CSI reporting on PUCCH supports Type I CSI. The semi-persistent CSI reporting on PUCCH format 2 supports Type I CSI with wide- band granularity, whereas semi-persistent CSI reporting on PUCCH format 3 or 4 supports Type I subband CSI and Type II CSI with wideband frequency granularity. When PUCCH carry Type I CSI with wideband frequency granularity, the CSI payloads carried by PUCCH format 2, and PUCCH format 3 or 4 are identical irrespective of RI and CRI. For New Radio Access Physical Layer Aspects (Part 2) 607 Type I CSI subband reporting on PUCCH format 3 or 4, the payload is split into

two parts. The first part may contain RI, CRI, and/or CQI for the first codeword. The second part con- tains PMI and the CQI for the second codeword when RI > 4. A semi-persistent report car- ried on PUCCH format 3 or 4 supports only part 1 of Type II CSI feedback. Supporting Type II CSI reporting on PUCCH format 3 or 4 is considered a UE capability. A Type II CSI report (part 1 only) is carried on PUCCH format 3 or 4 and is calculated independent of any Type II CSI reports carried on PUSCH. When the UE is configured with CSI report- ing on PUCCH format 2, 3, or 4, each PUCCH resource is configured for each candidate uplink BWP. The UE will never report CSI with a payload size larger than 115 bits when configured with PUCCH format 4 [8]. The NR physical uplink control channel is used for transmission of HARQ ACK/NACK for received downlink data; CSI feedback related to the downlink channel conditions to assist dynamic scheduling, and SR indicates that a UE needs uplink resources for data transmis- sion. The NR PUCCH supports two transmission modes (see Fig. 4.77): PUCCH format 0 and 2: number of PRBs = {1,2, ,16} PUCCH format 1,3,4: number of PRBs = PUCCH {1,2,3,4,5,6,8,9,10,12,15,16} PUCCH starting PRB PUCCH format 0 and 2: starting symbol = {0,1,...,13} PUCCH format 0 and 2: number of symbols = {1,2} PUCCH format 1,3,4: starting symbol = {0,1,...,10} PUCCH format 1,3,4: number of symbols = {4,5,...,14} Figure 4.77 General time-frequency structure of PUCCH for various formats [6]. 608 Chapter 4 Short-PUCCH: NR PUCCH can be transmitted in short duration around the last uplink symbol(s) of a slot. The transmission can span one or two OFDM symbols. Long-PUCCH: NR PUCCH can be transmitted in long duration at least over four uplink symbols to improve coverage. It is also considered to support a long-PUCCH transmis- sion over multiple slots. One or multiple PRBs can be allocated as the minimum resource unit size in frequency domain for short-PUCCH and long-PUCCH. Intra-slot frequency hopping can be configured for PUCCH

format 1, 3, or 4, that is, the long-PUCCH, where the number of symbols in the first is given by NPUCCH / 2 in which NPUCCH is the length of PUCCH transmission in the number of OFDM symbols. PUCCH formats 0, 1, 3, and 4 use the low-PAPR Zadoff-Chu sequences p(a)5)(n)|s=0 where a is a cyclic shift of a base sequence ru,v(n) such that = ejanru.v(n), 0n<Mzc in which Mzc = 12m is the length of the sequence (assuming 8 = 0). Multiple sequences can be generated from a single base sequence based on different values of a (cyclic shift). Base sequences (n) are divided into groups, where UE {0, 1, 29} is group number and V is the base sequence number within the group, such that each group con- tains one base sequence v=0 each of length Mc = 12m, 1 m5 and two base sequences v=0,1 each of length Mzc = 12m, mz6. The definition of the base sequence vu,v(0),..., depends on the sequence length Mc [6]. The sequence group (fgh + fss)mod 30 and the sequence number V within the group depend on the value of the RRC parameter pucch-GroupHopping. For example, if RRC parameter pucch-GroupHopping set to = fss = NID mod 30; v=0 where NID is the hopping identifier identified by RRC signaling, and c(n) is a pseudo-random sequence which was defined earlier. The frequency hopping index nhop = 0, if frequency hop- ping is disabled; otherwise, nhop = 0 for the first hop and nhop = 1 for the second hop. The cyclic shift a is a function of the symbol and slot number and is expressed as Al = ([mo + mcs + ncs(nslot,1+1)] mod 12)/6 where the parameters are defined in [6]. In the latter equation the function ncs 1) is defined as ncs(nslot, 1) = -2 :(8NSymb nslot + 8l + m) in which the pseudo-random sequence c(i) was defined earlier and the pseudo-random sequence generator is initialized with the RRC parameter hoppingId. The general time-frequency structure of PUCCH for various formats is depicted in Fig. 4.77. 4.2.3.1.1 PUCCH Format 0 Structure and Physical Processing PUCCH format 0 is a short PUCCH format that can transport up to 2 bits. It is use

d for HARQ-ACK and SRs. As shown in Fig. 4.78, PUCCH format 0 sequence is generated 11 and 1=0 for single-symbol PUCCH transmission and l =0,1 = for double-symbol PUCCH transmission. The sequence p(a,8, (n) was defined in the previous section, in which the parameter Mcs depends on the UCI. The sequence x(n) is New Radio Access Physical Layer Aspects (Part 2) 609 Scheduling request HARQ-ACK (1 or 2 bits) Phase rotation One OFDM symbol Per symbol pseudo random variation Configurable ID Per slot pseudo random variation Base sequence Base sequence One PRB (NACK/ACK) (NACK/ACK, -SR) (NACK/ACK, +SR) (NACK/NAC +SR) (ACK/ACK) (NACK/NACK)(ACK/ACK, -SR) (NACK/NACK, -SR) (ACK/ACK, +SR) Active uplink BWP (ACK/NACK, +SR) (ACK/NACK) (ACK/NACK, -SR) Figure 4.78 Time-frequency structure of PUCCH format 0 and signaling via phase rotation [6,14]. multiplied by amplitude scaling factor BPUCCH-FO in order to adjust the transmit power and is mapped sequentially to resource elements (k,l) assigned for PUCCH transmission in frequency-first manner on a single-antenna port [6]. As shown in Fig. 4.78, the phase rotations through parameter a represent different informa- tion bits that are separated with TT and /2 for 1- and 2-bit HARQ-ACK, respectively. In the case of a simultaneous SR, the phase rotation is increased by /4 for 1-bit acknowledgments and by /6 for 2-bit acknowledgments. The base sequences are configured per cell using an identity provided as part of the SI. Furthermore, a sequence hopping, where the base sequence varies on a slot-by-slot basis, can be used to randomize the interference between different cells. PUCCH format 0 is typically transmitted at the end of a slot. If two OFDM symbols are used for PUCCH format 0, the same information is transmitted on both OFDM symbols. However, the reference phase rotation as well as the frequency-domain resources may vary between the symbols, effectively creating a frequency-hopping effect [6,14]. 4.2.3.1.2 PUCCH Format 1 Structure and Physical Processing PUCCH format 1 is the long-

format version of PUCCH format 0, which can carry up to 2 bits, using 4-14 OFDM symbols over one resource block per symbol in frequency. The OFDM symbols are split between symbols used for control information and symbols desig- nated to the reference signals to enable coherent reception. The number of symbols used for control information and reference signals is a trade-off between channel estimation accuracy and energy in the information part. It was shown that a reasonable trade-off would be achieved if half of the symbols are used for reference signals. The block of bits b(0), b(NUCI 1)|Nuci =0,1 = is modulated using BPSK for a single-bit payload and 610 Chapter 4 QPSK for double-bit payload, resulting in a complex-valued symbol d(0). The complex- valued symbol d(0) is multiplied by a sequence such that y(n) = The block of complex-valued symbols y(0), y(11) is block-wise spread with the (DFT) orthogonal sequence wi(m) = exp(j2(m)/N5F) such that n=0,1,. , 11;m=0,1,.. NPUCCH-F SF,m' -1 where m' = 0 when there is no intra-slot frequency hopping and m'=0,1 when intra-slot hopping is enabled. The parameter NSUMCH-FI PUCCH-F1 SF,m and the orthogonal sequence wi(m) are defined in [6]. In case of a PUCCH transmission spanning multiple slots, the complex- valued symbol d(0) is repeated for the subsequent slots. The sequence z(n) is scaled with the scaling factor BPUCCH-F1 in order to adjust the transmit power and is mapped sequen- tially to resource elements (k,1) if they are not used by DM-RS. The mapping to resource elements (k,1) designated to PUCCH transmission is in increasing order of frequency and then time over the assigned physical resource block on a single-antenna port. The time- frequency structure of PUCCH format 1 is illustrated in Fig. 4.79. One or two UCI bits BPSK/QPSK modulation Cyclic shift Per symbol pseudo random variation Phase rotation Configurable ID Base sequence Per slot pseudo random variation Base sequence Length 4 orthogonal sequence DM-RS Length 5 orthogonal sequence Active uplink No frequ

ency hopping Frequency hopping Figure 4.79 Time-frequency structure of PUCCH format 1 [6]. New Radio Access Physical Layer Aspects (Part 2) 611 Unlike LTE, where PUCCH frequency-hopping was always done at the slot boundary, the NR provides additional flexibility by allowing variable PUCCH duration depending on the scheduling decisions and overall system configuration. Furthermore, since the devices are supposed to only transmit within their active bandwidth part, hopping is typically not between the edges of the overall carrier bandwidth as in LTE. Therefore, frequency hopping is configurable and determined as part of PUCCH resource configuration. The position of the hop is obtained from the length of PUCCH. If frequency hopping is enabled, one orthogonal block-spreading sequence is used per hop. 4.2.3.1.3 PUCCH Format 2 Structure and Physical Processing PUCCH format 2 is a short PUCCH format which is used to carry more than two uplink control bits, for example, CSI report and HARQ acknowledgments, or HARQ acknowledg- ments per se. An SR can also be included in the bits and jointly encoded. If the payload to be transmitted by PUCCH format 2 is too large, the CSI reports are not transmitted in order to make room for more important HARQ acknowledgment bits. The overall PUCCH format 2 physical layer processing is depicted in Fig. 4.80. A CRC is added for large payload sizes UCI bits Channel coding Scrambling sequence modulation PUCCH Scrambled allocated modulated and bandwidth encoded UCI bits -Slot- DM-RS Figure 4.80 Processing and mapping of PUCCH format 2 [6]. 612 Chapter 4 and the resulting block of bits are encoded using Reed-Muller - codes20 for payload sizes up to 12 bits or polar codes for larger payloads. The encoded block is scrambled and QPSK modulated. The scrambling sequence is based on the device identity (C-RNTI) together with the physical-layer cell identity (or a configurable virtual cell identity), ensuring inter- ference randomization across cells and devices using the same set of time-frequency r

esources. The QPSK-modulated symbols are then mapped to subcarriers across multiple resource blocks using one or two OFDM symbols. A pseudo-random QPSK sequence, mapped to every third subcarrier on each OFDM symbol, is used as a DM-RS to facilitate coherent detection at the base station [14]. The physical layer processing of UCI starts with CRC attachment and channel coding (see Fig. 4.80). Let us a1, aNuci-1 denote the input UCI bit sequence, in which NUCI is the payload size. If NUCI 12 the UCI bits are encoded with the polar codes and if NUCI VI 11, Reed-Muller codes, simplex code (NUCI = 2), or repetition coding (NUCI = 1) would be used to encode the UCI bits [7]. If the payload size NUCI 12, code block segmentation and CRC attachment may be performed prior to channel coding. If NUCI 360, E > 1088 bits or if NUCI > 1013, = 1 segmentation is performed; otherwise, Iseg = 0 and no segmentation is done, where E is the rate-matched output sequence length. If 12 VI NUCI < 19, the parity bits Pro,Pr1 Pr(NcRc-1) are computed by setting NCRC = 6 bits using the generator polynomial gCRC6(D). The resulting sequence is C10, C11, C1(K1-1) where l is the code block number and K is the number of bits associated with the 1th code block. If NUCI > 20, the parity bits are computed by setting NCRC = 11 bits using the generator polynomial gCRC11(D) [7]. Note that if the payload size NUCI < 11, no CRC bits are attached. The information bits are later encoded by channel coding block. The total number of code blocks is denoted by C and each code block is individually encoded. If 18 < K1 < 25, the information bits are encoded with polar encoder by setting the parameters as follows: nmax = 10, I = 0, NPC = 3, NPC = 1, if E - K1 + 3 > 192; otherwise, nwm=0, in which E1 is the rate-matched output sequence length of the 1th code block. If K1>30, the Reed-Muller codes are a family of linear error-correcting codes used in communication systems. The special cases of Reed-Muller codes include Hadamard codes, Walsh-Hadamard codes, and Reed-S

olomon codes. Reed-Muller codes are denoted by RM(d,r) notation, where d is the order of the code, and r determines the length of code n = 2r. Reed-Muller codes are related to binary functions on field GF(2') over the elements {0, 1}. It can be shown that RM(0,r) codes are repetition codes of length n = 2', rate R = 1/n, and minimum distance dmin = n, and RM(1,r) codes are parity check codes of length n = 2', rate R = (r + 1)/n, and mini- mum distance dmin = n/2. Reed-Muller codes have the following properties [15]: 1. The set of all possible exterior products of up to dof Vi form a basis for F2. The rank of RM(d,r) code is defined 3. RM(d,r) = RM(d,1 1)|RM(d- l' denotes the bar product of two codes. 4. RM(d,r) has minimum Hamming weight 2d-r. New Radio Access Physical Layer Aspects (Part 2) 613 Table 4.26: Rate-matched UCI output sequence length EUCI [7]. UCI(s) for Transmission on a PUCCH UCI for Encoding Value of EUCI HARQ-ACK HARQ-ACK Euci = ET HARQ-ACK, SR HARQ-ACK, SR EUCI = ET CSI (consisting of EUCI = ET one part) HARQ-ACK, CSI HARQ-ACK, CSI EUCI = ET (consisting of one part) HARQ-ACK, SR, HARQ-ACK, SR, EUCI = ET CSI (CSI not of two parts) CSI (consisting of CSI Part 1 two parts) CSI Part 2 HARQ-ACK, CSI HARQ-ACK, CSI Euci = (consisting of two Part 1 parts) CSI Part 2 EUCI = - HARQ-ACK, SR, HARQ-ACK, SR, Euci = CSI (consisting of CSI Part 1 two parts) CSI Part 2 information bits are encoded with polar encoder by setting the parameters as follows: nmax = 10, I = 0, NPC =3, nwm=0. The output bits following the encoding are denoted by do, d1,..., dN-1, where N is the number of coded bits. For PUCCH format 2/3/4, the total rate-matched output sequence length ET is given by Table 4.26, where NPUCCHEZ symb,UCI' Nsymb,UCI denote the number of symbols carry- ing UCI for PUCCH format 2/3/4;NPUCCH,NPUCCHare the number of PRBs that are determined by the UE for PUCCH format 2/3 transmission; and NPUCCH,4 spreading factor for PUCCH format 4 [7,8]. The rate matching is performed by setting IBIL = 1 and the rate matchin

g output sequence length to E = EUCI/CUCI where CUCI is the number of code blocks for UCI and the value of EUCI is given by Table 4.26. In this table the follow- ing parameters have been used [7]: OACK is the number of bits for HARQ-ACK for transmission on the current PUCCH. denotes the number of bits for SR for transmission on the current PUCCH. CSI-part is the number of bits for CSI part 1 for transmission on the current PUCCH. 2 denotes the number of bits for CSI part 2 for transmission on the current PUCCH. Rmax UCI is the configured PUCCH maximum coding rate. The output bit sequence after rate matching is denoted by where E is the length of rate-matched output sequence in 1th code block number. 614 Chapter 4 The encoded UCI payload b(0), b(NE - 1), where NE denotes the number of encoded bits in the payload transmitted on the physical uplink control channel, is scrambled prior to modulation, resulting in a block of scrambled bits b(0), ., b(NE - 1) such that b(i)=[b(i)+c(i)]: mod 2, in which the scrambling sequence c(i) is a pseudo-random sequence that is initialized with the value Cinit = NRNTI215 + nID that is derived from RRC configured parameter NID E {0, 1, 1023}; otherwise, NID The block of scrambled b(0) b(NE - 1) is QPSK modulated, resulting in a block of complex-valued modula- tion symbols d(0), d(NE/2-1). The block of modulation symbols is scaled with the scaling factor BPUCCH-F2 to adjust the transmit power and is sequentially mapped, starting with d(0), to resource elements (k,1) which are reserved for PUCCH transmission and are not used by the associated DM-RS. The mapping to resource elements (k,l) is in increasing order of the frequency index k followed by time index l on a single-antenna port. PUCCH format 2 is typically transmitted at the end of a slot as illustrated in Fig. 4.80; however, it is also possible to transmit PUCCH format 2 in other positions within a slot [6]. 4.2.3.1.4 PUCCH Formats 3 and 4 Structure and Physical Processing PUCCH format 3 is the long PUCCH counterpart to PUCCH f

ormat 2, wherein more than two UCI bits can be transmitted over 4-14 OFDM symbols, where there can be multiple resource blocks on each symbol. As a result, it is the PUCCH format with the largest pay- load capacity. The OFDM symbols are used for carrying the UCI and the PUCCH DM-RS. The control information is encoded using Reed-Muller codes for payload sizes less than 11 bits and polar codes for larger payloads and then scrambled and modulated. The scram- bling sequence is based on the UE identity (C-RNTI) together with the physical-layer cell identity (or a configurable virtual cell identity), ensuring interference randomization across cells and devices that use the same set of time-frequency resources. Prior to channel coding stage, a CRC is attached to the control information for large payloads. The encoded bits are QPSK modulated; however, there is an option to use 2-BPSK modulation to lower the cubic metric at the expense of some loss in link performance. The complex-valued modula- tion symbols are distributed among the OFDM symbols and DFT precoding is performed to reduce the cubic metric and improve the power amplifier efficiency. The structure of PUCCH format 4 is similar to that of PUCCH format 3 with the possibility to code-multiplex multiple devices over the same resources using one resource block in the frequency domain. Each OFDM symbol carries 12/NSF unique modulation symbols. Prior to DFT-precoding, each modulation symbol is block-spread with an orthogonal sequence of length NSF. The spreading factors of length two and four are supported, implying that the multiplexing capacity would be two or four devices on the same set of resource blocks. As shown in Fig. 4.81, frequency hopping can be configured for PUCCH format 3/4 to exploit frequency diversity; however, these PUCCH formats can operate without frequency hopping. The location of the reference signal symbols depends on the frequency hopping and the length New Radio Access Physical Layer Aspects (Part 2) 615 UCI bits Channel coding Scrambling se

quence modulation Multiple resource blocks DM-RS Active uplink DM-RS PUCCH format 3 UCI bits Channel coding Scrambling sequence modulation Length N block spreading per OFDM symbol DM-RS Active uplink DM-RS resource block PUCCH format 4 Figure 4.81 Short-PUCCH time-frequency structure 6,14]. 616 Chapter 4 of the PUCCH transmission, since at least one resource signal per hop is required. There is also a possibility to configure additional reference signal locations for longer PUCCH dura- tions to ensure two reference signal instances per hop. The UCI mapping is such that higher priority is given to HARQ-ACK, SR, and CSI part 1, which are jointly encoded and mapped around the DM-RS locations, and lower priority to the remaining bits which are mapped to the remaining positions [14]. More specifically, the encoded payload bits b(0) b(NE are scrambled prior to modu- lation, resulting in a block of scrambled bits b(0) b(i) = =b(i)+c(i))mod 2 where the scrambling sequence c(i) is a pseudo-random sequence that is initialized with Cinit = NRNTI215 + NID where parameter NID E {0,1,..., 1023} is deter- mined via RRC signaling; otherwise, it is set to the default value NID = Ncell The block of scrambled bits b(0),...,B(NE 1) is QPSK or 2-BPSK modulated, resulting in a block of complex-valued modulation symbols d(0), where for QPSK and = for 2-BPSK. For PUCCH formats 3 and 4, MPUCCH,s 12MRUCCH,S, in which MRUCCH,S denotes the bandwidth of the PUCCH in terms of the number of resource blocks. The parame- ter MPUCCH,S =2a23035a5 for PUCCH format 3 and MRUCCH,S = 1 for PUCCH format 4, wherein A3, a5 is a set of non-negative integers, and SE {3,4} identifies the format. For PUCCH format 3, no block-wise spreading is applied and y(IMPUCCH,3 +k) = where applied that(M where =0,1, Furthermore, PUCCH,4 E {2,4} and the spreading codes Wm(k) are given in [6,8]. The block of complex-valued symbols y(0), 1) is DFT transformed (precoded) as follows [6]: ..,MPUCCH,S resulting in a block of complex-valued symbols - 1), which are scaled with a

scaling factor BPUCCH,S to properly adjust the transmit power and are subse- quently mapped starting with z(0) to resource elements (k, 1) which are designated to PUCCH transmission and are not used by the associated PUCCH DM-RS. The mapping to New Radio Access Physical Layer Aspects (Part 2) 617 resource elements is done in frequency-first manner on a single-antenna port. In case of intra-slot frequency hopping, OFDM symbols are transmitted in the first hop and the in the second hop where NPUCCH,S symb the total number of OFDM symbols used in one slot for PUCCH transmission [6]. The physical processing and structure of PUCCH formats 3 and 4 are illustrated in Fig. 4.81. 4.2.3.2 Physical Random-Access Channel The NR supports a four-step random-access procedure similar to LTE. However, beam- forming and beam tracking aspects introduced in NR random-access procedure make the overall process different from that of LTE in frequencies above 6 GHz. The UEs need to detect and select the best beam for RACH process (beam selection process) prior to PRACH sequence selection and transmission. The PRACH design in NR relies on Zadoff-Chu sequences for the preamble construction. There are three PRACH preamble formats with long sequence length of 839, two of which, with subcarrier spacing of 1.25 kHz (same as LTE), are used for LTE refarming and large cells (up to 100 km); another format with subcarrier spacing of 5 kHz is defined for high-speed scenarios (up to 500 km/h) and cell radius up to 14 km (see Tables 4.27 and 4.28). Long sequences support Table 4.27: PRACH preamble formats for LRA = 839 and Afra E {1.25,5} kHz [6]. PRACH Subcarrier Spacing Bandwidth Support for Format Afra (kHz) (MHz) Restricted Sets 24576Ts 3168Ts 2976Ts Type A, Type B 24576Ts 21024Ts 21984Ts 24576Ts 4688Ts 29264Ts 24576T 3168Ts 2976Ts Table 4.28: Preamble formats for LRA = 139 and Afra = 2H X 15 kHz where HE {0, 1,2,3} and K = 64 [6]. PRACH Format LRA Samples Afra (kHz) NRA Samples NRA Samples 2 th X 15 2048Ts 2 X 2048k2 -" 288k2-H 4 X 2048k2 - 57

6k2-r 5 X 2048k2 - 864k2-11 2 2X2048k2-" 216k2-H 4 X 2048k2 -4 3602-u 6 X X 2048k2 - 504k2- 12 X 2048k2 - 936k2- 1 X2048k2 12402-H 4 X2048k21 20482- 618 Chapter 4 unrestricted sets and restricted sets of Type A and Type B, while short sequences support unrestricted sets only. Considering network beam-sweeping reception within a RACH occasion, NR introduced a new set of PRACH preamble formats of shorter sequence length of 139 on 1, 2, 4, 6, and 12 OFDM symbols and subcarrier spacings of 15, 30, 60, and 120 kHz. The new formats are composed of single or consecutive repeated RACH sequences. The cyclic prefix is inserted at the beginning of the preambles, and the guard time is appended at the end of the preambles, while the cyclic prefix and gap between RACH sequences is omitted. For both short and long PRACH preamble sequences, the network can also conduct beam-sweeping reception between RACH occasions. Multiple RACH preamble formats are defined with one or more PRACH symbols, and dif- ferent cyclic prefix and guard time lengths. The PRACH preamble configuration is signaled to the UE in the SI. The UE calculates the PRACH transmit power for the retransmission of the preamble based on the most recent estimate of pathloss and power ramping counter. If the UE conducts beam switching, the counter of power ramping remains unchanged. The SI informs the UE of the association between the SS blocks and the RACH resources. The threshold of the SS block for RACH resource association is based on the RSRP and is net- work configurable. Prior to initiation of the physical random-access procedure, the physical layer of the UE must receive a set of SS/PBCH block indices and provide the UE RRC sublayer with the corresponding set of RSRP measurements conducted on those SS/PBCH candidates. The information required for the UE physical layer prior to PRACH transmission includes PRACH preamble format, time resources, and frequency resources for PRACH transmission as well as the parameters for determining the root sequences and their cycl

ic shifts in the PRACH preamble sequence set including index to logical root sequence table, cyclic shift Ncs, and set type, that is, unrestricted, restricted set A, or restricted set B. Physical random-access procedure is triggered following a request from the UE RRC sub- layer or by a PDCCH command (control element) with the following parameters [8]: con- figuration for PRACH transmission, preamble index, preamble subcarrier spacing, transmit power target PRACH' RA-RNTI, and a PRACH resource. The PRACH preamble is transmitted using the selected PRACH format with transmission power value over the designated PRACH resource. The SS/PBCH block indices are mapped to PRACH occasions (see Fig. 4.82); first, in increasing order of preamble indices within a single PRACH occasion followed by, in increasing order of frequency resource indices of frequency-multiplexed PRACH occasions, then, in increasing order of time resource indices of time-multiplexed PRACH occasions within a PRACH slot and, finally, in increasing order of indices of PRACH slots. The associ- ation period, starting from frame 0, for the mapping of SS/PBCH blocks to PRACH occasions New Radio Access Physical Layer Aspects (Part 2) 619 PRACH slot PRACH slot PRACH slot PRACH slot PRACH-Preamble PRACH PRACH configuration period SSBO SSB 1 SSB 2 SSB 3 Rach-Preamble PRACH SSB SSB 1 SSB 2 SSB 3 PRACH slot i PRACH slot it One PRACH occasion Figure 4.82 Structure of NR PRACH opportunities in time and frequency domain 14]. Table 4.29: Mapping between PRACH configuration period and SS/PBCH block to PRACH occasion association period [8]. PRACH Configuration Association Period (Number of PRACH Period (ms) Configuration Periods) {1,2,4,8,16} {1,2,4,8} {1,2,4} {1,2} is the smallest value in a set (see Table 4.29) determined by the PRACH configuration period such that NSSB TX SS/PBCH blocks are mapped at least once to the PRACH occasions within the association period. A UE obtains the parameter NSSB TX from SystemInformationBlockType1. If after an integer number of SS/PB

CH blocks to PRACH occasions mapping cycles within the association period, there is a set of PRACH occasions that are not mapped to N TX SSB SS/PBCH blocks; no SS/PBCH blocks are mapped to the set of PRACH occasions. An association pat- tern period includes one or more association periods and is calculated such that a pattern between PRACH occasions and SS/PBCH blocks repeats at most every 160 ms. The PRACH 620 Chapter 4 occasions that are not associated with SS/PBCH blocks after an integer number of association periods, if any, are not used for PRACH transmissions [8]. If a random-access procedure is initiated by a PDCCH command, the UE must transmit the PRACH preamble in the first available PRACH occasion for which the time interval between the last symbol of the PDCCH reception and the first symbol of the PRACH transmission is larger than or equal to TN2 + ABWP-Switching + Delay (in milliseconds), where TN2 is the equivalent time duration of N2 symbols corresponding to PUSCH preparation time assuming certain PUSCH processing capability, the parameter WP-Switching = 0, if uplink active BWP does not change, and A Delay = 0.5 ms for FR1 and Delay = 0.25 ms for FR2 [8]. The PRACH preamble transmission can occur within a configurable subset of slots known as the PRACH slots (see Fig. 4.82) that are repeated every PRACH configuration period. There may be multiple PRACH occasions within each PRACH slot in the frequency-domain N jPRACH-Preamble RB NPRACH consecutive resource blocks where N PRACH-Preamble cover preamble bandwidth measured in number of resource blocks and NPRACH is the number of frequency-domain PRACH occasions. For a given preamble type, corresponding to a certain preamble bandwidth, the overall available time-frequency PRACH resources within a cell can be described by the following parameters: a configurable PRACH periodicity that can range from 10 to 160 ms; a configurable set of PRACH slots within the PRACH period; a configurable frequency-domain PRACH resource given by the index of the first resour

ce block in the resource and the number of frequency-domain PRACH occasions [14]. In NR, the set of random-access preambles xu,v(n) is generated based on Zadoff-Chu sequences such that xu,v(n) = =xx[(n+C,)mod LRA] wherein Xu(i) = exp( - jui(i + 1)/LRA) is a Zadoff-Chu sequence of length LRA and root index ; and i = 0, 1, LRA - The frequency-domain representation of the preamble sequence is obtained by taking LRA - point DFT of the sequence Xu,v(n), resulting in = exp( - j2mn/LRA) where the sequence length LRA = 839 or LRA = 139 depend on the PRACH preamble format (see Tables 4.27 and 4.28). There are 64 preambles in each time-- frequency PRACH occasion, numbered in increasing order of cyclic shift Cv of a logical root sequence and increasing order of the logical root sequence index, starting with the index obtained via RRC signaling. The sequence number is obtained from the logical root sequence index [6]. The output of the DFT is then repeated NSEQ times, after which a cyclic prefix is inserted. For the PRACH preamble, the cyclic prefix is not inserted per OFDM symbol, rather it is inserted only once for the block of NSEQ repeated symbols (see Fig. 4.83). New Radio Access Physical Layer Aspects (Part 2) 621 Format 0 Preamble sequence fra 1.25 kHz NSEQ = 1 NRA A Format 1 Preamble sequence Preamble sequence fra 1.25 kHz NSEQ = 2 Format 3 fra = 5 kHz NSEQ = 4 NRA H General structure Preamble Preamble Preamble of short PRACH sequence 0 sequence 1 sequence NSEQ-1 preambles NRA H repetition of the preamble sequence) Figure 4.83 NR PRACH preamble structures [6]. The time-domain representation of PRACH signal SPRACH(t,I) on a single-antenna port is given as follows [6]: Chapter 4 Table 4.30: Supported combinations of Afra and Af, and k [6]. Afra for PRACH for PUSCH NRA E (RBs) for PUSCH In the preceding expression k is given in Table 4.30 [6]. Af is the subcarrier spacing of the active uplink bandwidth part during the initial access; otherwise, If is the subcarrier spacing of the active uplink bandwidth part (see Table

4.30 for permissible values). BWP,i is the lowest numbered resource block of the initial active uplink bandwidth part based on common resource block indexing and is derived via RRC parameter initialUplinkBWP during initial access; otherwise, Nstart BWP,i is the lowest numbered resource block of the active uplink bandwidth part based on common resource block indexing and is derived by the higher layer parameter BWP-Uplink. nstart NRA is the frequency offset of the lowest PRACH transmission occasion in the frequency-domain relative to PRB_0 of the initial active uplink bandwidth part given by the RRC parameter msgl-FrequencyStart during initial access associated with the initial active uplink bandwidth part; otherwise, nstart is the frequency offset of lowest PRACH transmission occasion in frequency domain with respect to physical resource block 0 of the active uplink bandwidth part given by the RRC parameter prach- frequency-start associated with the active uplink bandwidth part. NRA is the PRACH transmission occasion index in the frequency-domain for a given PRACH transmission occasion in time. RB is the number of resource blocks that are occupied by PRACH preamble. The starting position of PRACH preamble in a subframe RA start Afra E {1.25, 5, 15, 30} kHz or in a slot with 60 kHz subcarrier spacing when New Radio Access Physical Layer Aspects (Part 2) 623 defined start l = 0; otherwise, start start,1- + The subframe or 60 kHz slot is assumed to start at t = 0. The timing advance is assumed to be zero NTA = 0. The numerology corresponding to Afra E {1.25,5} kHz is assumed to be u = 0; otherwise, it is given by Afra E 60, 120} kHz, and the symbol position l is given by = lo + + 14nRA where lo is the starting symbol, nRA is the PRACH transmission occasion within the PRACH slot, num- bered in increasing order from 0 to NRA,slot - 1 within a RACH slot, NRACtion duration is given and the nRA depends on fra, that is, if fRA E {1.25,5 15, 60} kHz then nRA = 0; otherwise if {fra e{30,120} kHz and either of the number of

PRACH slots within a subframe or num- ber of PRACH slots within a 60 kHz slot is equal to 1, then = 1; otherwise, 0,1 The quantities LRA and Nu are the length of the PRACH sequence and the number of sam- ples in a PRACH symbol, and + 16kn wherein n = 0 for Afra E {1.25,5} kHz. For AfRAE{15,30,60,120}kHz, n is the number of times the interval overlaps with either time instance zero or time instance (AfmaxN//2000)T, =0.5 ms in a subframe [6]. The parameters ZeroCorrelationZoneConfig and prach-RootSequenceIndex (defined in [6]) are used to generate the random-access signatures for each cell, which are required to be distinct across neighboring cells. There is a relationship between the preamble format and the cell radius, which means that the selection of ZeroCorrelationZoneConfig parameter is related to the cell radius. The parameters ZeroCorrelationZoneConfig and prach-RootSequenceIndex are derived from SystemInformationBlockTypel. The random- access sequences are generated via selection of a Zadoff-Chu sequence (1 out of 839 or 139) given by prach-RootSequenceIndex and a cyclic shift that is used 64 times to generate the 64 random-access signatures from the Zadoff-Chu sequence selected. The cyclic shift is indirectly provided to the UE via the parameter ZeroCorrelationZoneConfig. The cyclic shift is also related to the cell size. The relationship between the cyclic shift and the cell size is given by (Ncs - 1) (800 us/839) > RTD + TDelay_Spread. If Afra = 1.25 kHz, the PRACH symbol duration is 0.8 ms (0.133 ms in case of 139). The round-trip delay can be written as RTD = 2Rcell/c; therefore Rcell VI c[(Ncs 1)(800 us/839) - T Delay_Spread]/2 As an example, if we assume that ZeroCorrelationZoneConfig is 12 then from [6] and assum- ing Afra = 1.25 kHz, NCS = 119. Furthermore, if TDelay_Spread = 6 us then the cell size will be approximately 15.97 km. Note that the smaller the cyclic shift, the smaller cell size. The delay spread in the preceding expression is derived empirically and the value of the delay spread is t

ypically different for rural, suburban, urban and dense urban environments. In practice, the ZeroCorrelationZoneConfig parameter is a pointer to a table that provides a set of available cyclic shifts in the cell, where different tables indicated by this parameter have different distances between the cyclic shifts, thus providing larger or smaller zones or tim- ing errors for which orthogonality or zero correlation can be maintained. Chapter 4 Table 4.31: Random-access configurations for TDD mode in FR1 [6]. PRACH Preamble nSFN mod X = =y Subframe Number Starting Number PRACH,slot duration Configuration Format Symbol Number of PRACH Index PRACH Time-Domain Duration Slots PRACH Within a Occasions Subframe Within a RACH Slot 3,4,8,9 0,1,2,3,4,5,6,7,8,9 1,3,5,7,9 The PRACH preamble sequence is mapped to physical resources such that ak = BPRACHYu,v(k); k=0,1, LRA - 1 where BPRACH is a transmit power adjustment (p,RA) scaling factor, p is the antenna port from which the PRACH is transmitted. The PRACH preambles can only be transmitted in the time resources that are signaled via RRC parame- ter prach-ConfigurationIndex and further depend on frequency range FR1 or FR2 where the system is deployed and the spectrum type. The PRACH preambles can only be transmitted in the frequency resources specified by parameter msgl-FrequencyStart. The PRACH fre- quency resources NRA E {0, 1, M - 1}, in which the parameter M is derived from the RRC parameter msgl-FDM, are numbered in increasing order within the initial active uplink bandwidth part during initial access, starting from the lowest frequency. For the purpose of slot numbering, it is assumed that subcarrier spacing is 15 kHz for FR1 and 60 kHz for FR2. Table 4.31 provides random-access configurations for TDD mode in FR1 [6,8]. The transmission power for PRACH on the kth uplink BWP of the 1th carrier based on a certain SS/PBCH block determination for the mth serving cell in transmission period i is determined by the UE as PRACH(i) = min [PcMAX(i), plm PRACH-target + PLklm] (dB

m), wherein is the configured UE transmission power for the 1th carrier in the New Radio Access Physical Layer Aspects (Part 2) 625 mth serving cell within transmission period i, plm PRACH-target is the PRACH preamble target reception power PREAMBLE_RECEIVED_TARGET_POWER signaled via RRC for the kth uplink BWP on 1th carrier in the mth serving cell, and PLklm is the calculated pathloss for the kth uplink BWP corresponding to the 1th carrier for the current SS/PBCH block in the mth serving cell calculated by the UE in decibels. If within a random-access response win- dow, the UE cannot receive a random-access response that contains a preamble identifier corresponding to the preamble sequence transmitted by the UE, the UE will typically ramp up (in steps) the transmission power up to a certain limit for the subsequent PRACH trans- missions. If prior to PRACH retransmission, the UE changes the spatial domain transmis- sion filter; the physical layer will notify the higher layers to suspend the power ramping counter [8]. The random-access preamble sequence can be generated at the system sampling rate, by means of a large IDFT unit. The cyclic shift can be implemented either in the time domain after the IDFT or in the frequency domain before the IDFT through a phase shift. For all possible system sampling rates, both cyclic prefix and sequence duration correspond to an integer number of samples. The method of Fig. 4.84 does not require any time-domain fil- tering in the baseband but requires large IDFT sizes (up to 24576 for a 20 MHz spectrum allocation), which are practically prohibitive. Therefore, another option for generating the PRACH preamble consists of using small-sized IDFT and shifting the preamble to the required frequency location through time-domain up-sampling and filtering (hybrid fre- quency/time-domain generation). Assuming that the preamble sequence length is 839, the smallest IFFT size that can be used is 1024. The sizes of the random-access cyclic prefix and preamble sequence duration have been cho

sen to provide an integer number of samples at the system sampling rate. The cyclic prefix can be inserted before the up-sampling and time-domain frequency shift, in order to minimize the intermediate storage requirements. Assuming sampling rate of 30.72 MHz and considering that the random-access preamble spans 0.8 ms, it can be concluded that the number of samples in time is equal to 24576. Furthermore, if the PRACH subcarrier spacing is assumed to be 1.25 kHz and the subcarrier spacing for PUSCH and PUCCH is 15 kHz, in order to maintain the same sampling rate, a 12 X 2048 point DFT operation would be needed for the PRACH signal generation at the transmitter side, if the entire processing is done in the frequency domain. An alternative approach is to use time-domain signal generation and extraction which involves up- sampling and filtering operations at the transmitter. The drawback of time-domain imple- mentation is that the up-sampling from 1.08 MHz to the system sampling-rate of 30.72 MHz is difficult to implement. The implementation of the PRACH signal at the gNB receiver can take a frequency-domain or a hybrid time/frequency-domain approach. As illustrated in Fig. 4.84 as an example, the common parts to both approaches are the cyclic prefix removal, which always occurs at the Chapter 4 Sampling rate R Sampling rate fs To RF NR -point Subcarrier Repeat Cyclic prefix circuitry sequence Insertion mapping Preamble length T (N A samples) (SizeN PRACH transmitter architecture 1 Sampling rate fIFFT Sampling rate fs To RF Repeat Cyclic prefix Time- circuitry insertion sequence domain (Size NRA) sampling Phase shift PRACH transmitter architecture 2 Frequency domain method NDFT-point Subcarrier Concatenated periodic correlations of demapping all cyclically-shifted Versions of the Root ZC sequence Sampling rate f Sampling rate R Cyclic prefix Signature detection Received signal removal (Size NDFT +NCP) Time- domain NFFT-point Subcarrier Complex frequency demapping conjugation shift Signature0 Signature3 Signature1 Sig

nature2 N - Point Root ZC PRACH receiver architecture 1 sequence Sampling rate fIFFT Hybrid time-frequency method Power delay profile calculation Cyclic Subcarrier Linear Insert Decimation prefix demapping Energy filter zeros detection removal Numerically controlled Root ZC oscillator sequence PRACH receiver architecture 2 Figure 4.84 Example PRACH transmitter and receiver structure 15]. front-end at the system sampling rate, the power delay profile calculation, and signature detection. The two approaches differ only in the computation of the subcarriers carrying the PRACH signal(s). The frequency-domain method computes the full range of subcarriers used for uplink transmission over the system bandwidth from 0.8 ms-long received input samples. As a result, the PRACH subcarriers are directly extracted from the set of uplink subcarriers, which does not require any frequency shift or time-domain filtering but involves an extremely large DFT computation. Note that even though DFT size NDFT = n2m, and we New Radio Access Physical Layer Aspects (Part 2) 627 can use fast and effcient DFT computation algorithms, the DFT computation cannot start until the complete sequence is stored in memory, which increases the processing delay. 4.2.3.2.1 Four-Step Random-Access Procedure From the UE physical layer perspective, the RACH procedure consists of transmission of random-access preamble (Msg1) in a PRACH occasion, receiving random-access response message via PDCCH/PDSCH (Msg2), and transmission of Msg3 in PUSCH, and receiving PDSCH for contention resolution. If the random-access procedure is initiated by a PDCCH command, the PRACH preamble is transmitted with the same subcarrier spacing. The random-access procedure comprises four steps. However, before the UE can attempt to access the network, it must synchronize to the downlink and receive the SI via PBCH and PDCCH/PDSCH. Upon receiving the SI, the UE would have the knowledge of PRACH con- figuration and transmission parameters such as PRACH preamble format, time-frequency re

sources to transmit PRACH, the parameters for determining the root sequences and their cyclic shifts in the PRACH preamble sequence set, index to the logical root sequence table, cyclic shifts, and the associated set type, that is, unrestricted, restricted Type A, or restricted Type B [6,8,12]. More specifically, the RACH procedure consists of the following 4 steps [11,12]: In the first step, the UE transmits a PRACH preamble associated with an RA-RNTI, if all conditions for PRACH transmission are met [12]. The gNB calculates the RA-RNTI associ- ated with the PRACH occasion, in which the random-access preamble is transmitted, as fol- lows RA-RNTI = 1 + Sid + 14tid + 14 X 80fid + 14 X 80 X 8ulcarrier_id where 0 < 14 is the index of the first OFDM symbol of the specified PRACH;0 tid <80 denotes the index of the first slot symbol of the specified PRACH in a system frame;0 < fid <8 is the index of the specified PRACH in the frequency domain; and ulcarrier_id is the uplink carrier used for Msgl transmission (ulcarrier_id =0: NR uplink carrier, ulcarrier_id = 1: SUL carrier). The frequency-domain location (resource) for PRACH preamble is determined by the RRC parameter msgl-FDM and msgl-FrequencyStart. The time-domain location (resource) for PRACH preamble is determined by the RRC parameter prach-ConfigurationIndex. In the second step, following the PRACH transmission, the UE awaits random-access response from the gNB which would be sent through a DCI scrambled with RA-RNTI value calculated as above. The UE attempts to detect a PDCCH with the corresponding RA-RNTI within the period of ra-Response Window. The UE searches for the DCI in the Type 1 PDCCH common search space. The DCI format for scheduling RAR message on PDSCH DCI format 1_0 scrambled with RA-RNTI. The resource allocation type for the Msg2 on PDSCH is resource allocation Type 1. The frequency-domain resource allocation for the PDSCH carrying RAR message is specified by DCI format 0_1. The time-domain resource allocation for the RAR message on PDSCH is speci

fied by DCI format 1 and PDSCH- ConfigCommon. The RAR window is configured by rar-WindowLength information element 628 Chapter 4 in a SIB message. If the UE successfully detects the PDCCH, it can decode PDSCH carry- ing the RAR message. After decoding the RAR message, the UE checks, if the random- access preamble ID (RAPID) in the RAR message matches the RAPID assigned to the UE. The PDCCH and PDSCH associated with the process are expected to use the same subcar- rier spacing and cyclic prefix as SIB1. Note that the gNB is not expecting any HARQ-ACK for the RAR message. The gNB may conclude that UE has successfully received and decoded the RAR message, if the UE does not retransmit PRACH, which would happen if the UE does not detect the DCI format 1_0 with CRC scrambled with the corresponding RA-RNTI within the RAR window, or if the UE does not correctly receive the transport block in the corresponding PDSCH within that window. In the third step, the UE must determine whether it should apply transform precoding for Msg3 on PUSCH, based on the RRC parameter msg3-transformPrecoder. The UE deter- mines the subcarrier spacing for Msg3 on PUSCH based on the RRC parameter SubcarrierSpacing in BWP-UplinkCommon. The UE then transmits Msg3 on PUSCH to the same serving cell to which it had sent the PRACH. In the fourth step, Msg4 is transmitted to the UE for contention resolution. The UE starts ra-ContentionResolutionTimer and monitors PDCCH with TC-RNTI while ra-ContentionResolutionTimer is running. The UE looks for the DCI in Type 1 PDCCH common search space. If the PDCCH is successfully detected, the UE proceed to decode PDSCH carrying the MAC control element, and at the same time, it sets the value of the C- RNTI to TC-RNTI and discards ra-ContentionResolutionTimer. The UE considers the RACH procedure as successfully completed. Once the UE successfully decodes Msg4 (con- tention resolution), it sends HARQ-ACK for the data (PDSCH carrying Msg4). In response to the PDSCH reception with the UE contention resolution identi

SS block 4 Uplink 1 Uplink 2 Uplink 3 Uplink 4 ((H)I)) PRACH transmission (same PSS, SSS, PBCH reception transmit beam direction as in the downlink transmit beam) UE TX PUSCH RAR window NT1+0.5 ms UE RX PDSCH PDSCH Slot in Slot n+k2+ NT1+NT2+0.5 ms NT1 A time duration of N1 symbols related to PDSCH reception time for UE Last symbol of First symbol of Msg2 on PDSCH Msg3 on PUSCH processing capability 1 when additional PDSCH DM-RS is configured. NT2: A time duration of N2 symbols related to PUSCH preparation time for UE processing capability 1. Figure 4.85 Random-access procedure 8,12]. 630 Chapter 4 gNB UE (1) Random access preamble (1) MsgA: PRACH preamble + data (Msg1 + Msg3) (2) Random access response (2) MsgB: random-access response (Msg2 + Msg4) (3) Msg3 transmission on PUSCH (4) Contention resolution Figure 4.86 Comparison of two-step and four-step contention-based random-access procedures [11]. In NR Rel-15, the RACH procedure is triggered, when uplink data becomes available at the UE buffer and the UE is either in the RRC_IDLE/INACTIVE state, where the RACH pro- cedure is triggered for state transition, or in the RRC_CONNECTED state, if the uplink is not synchronized, where the RACH procedure is used to reestablish uplink synchronization, or in the RRC_CONNECTED state, if the UE has no PUCCH resources available for SR or the SR procedure fails, where the RACH procedure serves as an SR. In addition, the RACH procedure is used for beam failure and recovery, on-demand SI request, or it can be explic- itly triggered by the network with RRC for handover. In NR Rel-15, the uplink data cannot be transmitted until the RACH procedure is success- fully completed. It is observed that for small packet transmission, four-step RACH is not efficient in terms of latency and signaling overhead; thus the two-step RACH has been pro- posed to simplify the RACH procedure to achieve lower signaling overhead and latency. It is possible to allow Msg3 in four-step RACH procedure to carry data in order to reduce the latency and ove

rhead. However, even in that case, the four-step RACH would still involve more signaling and latency relative to that of the two-step approach. In two-step RACH, the MsgA may consist of two parts, that is, PRACH preamble and PUSCH which are time-division multiplexed. The PRACH preamble is used for UE detec- tion, allowing the network to prepare for the reception of the corresponding PUSCH mes- sage. In NR Rel-15, up to 64 preamble signatures are mapped to one PRACH occasion. The preambles are orthogonal, or quasi-orthogonal, allowing the network to receive multiple preambles (from different UEs) in the same PRACH occasion. If all the preambles are mapped to a PUSCH in the same time-frequency occasion and more than one preamble is detected, the PUSCH transmissions of the detected preambles overlap in time and fre- quency, increasing the probability of PUSCH decoding failure. Alternatively, each pream- ble, or subset of preambles, can be mapped to a PUSCH in a unique time-frequency resource. This reduces the probability of PUSCH decoding failure due to collision but sig- nificantly increases the two-step RACH physical-layer overhead in the uplink. The MsgB in New Radio Access Physical Layer Aspects (Part 2) 631 two-step procedure comprises several fields including the detected unique ID for contention resolution, where the size of the detected ID depends on the use case; a timing advance field; an uplink-grant for scheduling the data packets after the RACH procedure; and small user-plane/control-plane packets for downlink communication. The presence and the size of each field depend on the use case; thus the total size of MsgB may vary. 4.2.4 Physical Uplink Shared Channel The physical uplink shared channel is used to transmit the user traffic and control informa- tion in the uplink. It supports two transmission modes namely codebook-based and non- codebook-based multi-antenna transmission. For codebook-based transmission, the gNB provides the UE with a transmit precoding matrix indication in DCI. The UE uses the i

ndi- cator to select the PUSCH transmit precoder from a set of codebooks. For non-codebook- based transmission, the UE determines its PUSCH precoder based on (wideband) SRS resource indication (SRI) field from DCI. A closed-loop DM-RS-based spatial multiplexing is supported for PUSCH with up to four transmission layers for SU-MIMO with CP-OFDM waveform. Uplink SU-MIMO uses one codeword. Support of DFT-S-OFDM in the uplink is optional, and when transform precoding is used, only a single MIMO transmission layer is supported. As shown in Fig. 4.87, the uplink physical layer processing of transport channels consists of the following stages: Transport block CRC attachment, where TBSs larger than 3824 use 24-bit CRC and other TBSs utilize 16-bit CRC, followed by LDPC base graph selection Code block segmentation and code block CRC attachment, which always uses 24-bit CRC Channel coding which makes use of LDPC coding (base graph 1 or 2) Rate matching and code block concatenation followed by data and control multiplexing Scrambling and modulation where any of the modulation schemes may be used, that is, 2-BPSK (with transform precoding only), QPSK, 16QAM, 64QAM, or 256QAM Layer mapping Transform precoding (enabled/disabled by configuration) and precoding Mapping to assigned resources and antenna ports The UE transmits at least one symbol with DM-RS on each layer in which PUSCH is trans- mitted. The number of DM-RS symbols and resource element mapping is configured via RRC signaling. The PT-RS may be transmitted on additional symbols to assist the gNB receiver with phase tracking [11]. Following the above summary, let us discuss the physical layer processing of the UL-SCH in more detail. The CRC is calculated over the entire transport block that is constructed Chapter 4 PUSCH data (transport block) ao, a b(cw) Transport block CRC attachment Scrambling TS 38.212 Section 6.2.1 TS 38.211 Section 6.3.1.1 bo b bB-1 LDPC base graph selection Modulation TS 38.212 Section 6.2.2 TS 38.211 Section 6.3.1.2 Code block segmentation and

Code block CRC attachment Layer Mapping TS 38.211 Section 6.3.1.3 TS 38.212 Section 6.2.3 Channel coding Transform precoding TS 38.212 Section 6.2.4 TS 38.211 Section 6.3.1.4 (IN PUSCH Rate matching Precoding TS 38.212 Section 6.2.5 TS 38.211 Section 6.3.1.5 "-"(i)]' Code block concatenation Resource Mapping TS 38.212 Section 6.2.6 Mapping to VRB TS 38.211 Section 6.3.1.6 go, 81, Resource mapping Data and control multiplexing VRB-to-PRB mapping TS 38.212 Section 6.2.7 TS 38.211 Section 6.3.1.7 Figure 4.87 Physical processing of the uplink shared channel 30]. from the MAC PDU(s). We denote the TB bits by ao, a1, ANPUSCH and the parity Po, p1, PNCRC-1, where NPUSCH is the payload size and NCRC is the number of parity bits. The number of parity bits depends on the PUSCH payload size. If NPUSCH > 3824, NCRC CRC bits are computed and attached to the TB using the generator polynomial = D24 + D23 +D18+D17 +D14 + D11 +D10+ D7 + D6 + D5 + D4 + D3 otherwise NCRC = 16 CRC bits are calculated using the generator polynomial gCRC16(D) + D5 + 1. The code block bits after CRC attachment are denoted by , bNPUSCH+NCRC-1 [7]. For initial transmission of a TB with coding rate R, determined by the MCS index and the subsequent retransmissions of the same TB, each code block of the TB is encoded with either LDPC base graph 1 or 2 depending on the values of NPUSCH and R. If NPUSCH < 292 or if NPUSCH < 3824 and R < 0.67 or if R < 0.25, LDPC base graph 2 is used; otherwise, LDPC base graph 1 is used [7]. New Radio Access Physical Layer Aspects (Part 2) 633 The input bit sequence to the code block segmentation is denoted by bo, b1 , NNPUSCH+NCRC~1 If B = NPUSCH + NCRC is larger than the maximum code block size Kcb, the bit sequence is seg- mented, and a 24-bit CRC is attached to each (segmented) code block. The value of Kch for LDPC base graph 1 is Kcb = 8448 and for LDPC base graph 2, Kcb = 3840. The number of seg- mented code blocks is determined by C = The output bits from code block segmentation are denoted by C10, C11 C1(K1-1) where

0<1<C is the code block number, and K1 = K is the number of bits in the 1th code block. The code blocks are then delivered to the channel coding unit where each code block is individually encoded with the LDPC encoder. The encoded bits are denoted by du d(N,-1), in which N1 = 66Z- for LDPC base graph 1 and N = 50Z for LDPC base graph 2, where the value of the lifting factor Z is given in Table 4.7. The encoded bits for each code block are processed through the rate matching function. The total number of code blocks is denoted by C, and each code block is individually rate matched by setting ILBRM = 1, if RRC parameter rateMatching is set to limitedBufferRM; otherwise, by setting ILBRM 0. After the rate matching stage, the bits are denoted by ,fi(E1-1), where Er is the number of rate matched bits in the 1th code block. The input bit sequence to the code block concatenation module are the sequences {fik|l = 0, 1, 1; = 0, 1, E - 1} where E is the number of rate-matched bits in the 1th code block. The output bit sequence from the code block concatenation function is the sequence go, g1, gG-1 where G is the total number of coded bits for transmission. The code block concatenation function sequentially concatenate different rate-matched code blocks [7]. The NR supports UCI multiplexing on PUSCH when the UCI and PUSCH transmissions coincide in time, either due to transmission of an uplink TB or due to triggering of aperiodic CSI transmission without an uplink TB. The UCI carrying HARQ-ACK feedback with 1 or 2 bits is multiplexed by puncturing PUSCH. In all other cases, the UCI is multiplexed by rate matching PUSCH. In case of SUL, the UE is configured with two uplink carriers and one downlink carrier in the same cell. The uplink transmissions on the uplink carriers are controlled by the network to avoid overlapping PUSCH/PUCCH transmission in time- domain. Overlapping transmissions on PUSCH are avoided through scheduling, while over- lapping transmissions on PUCCH are avoided through configuration since PUCCH can only b

e configured for one of the two uplink carriers in the cell. The initial access is supported on each uplink carrier. The mapping of the UCI to PUSCH resources is such that more (operationally) important bits (HARQ-ACK) are mapped to the first OFDM symbol after the first DM-RS, and less (operationally) important bits (CSI reports) are mapped to the subsequent symbols. Unlike the data part, which relies on rate adaptation to overcome the effects of radio propagation, the L1/L2 control signaling part cannot be rate-adapted. Power control may theoretically be used, but it would imply fast power variations in the time 634 Chapter 4 domain, which negatively impact the RF properties. Therefore, the transmission power is maintained constant over the PUSCH duration and the amount of resource elements allo- cated to L1/L2 control signaling is changed by changing the code rate of the control signal- ing. In addition to a semi-static value controlling the amount of PUSCH resources used for UCI, it is also possible to signal this information as part of a DCI. The bit sequence from the output of code block concatena- tion and multiplexing, where is the number of bits in codeword q transmitted on the physical shared channel, are scrambled prior to modulation, resulting in a block of scram- bled bits (N(9) - such that the UL-SCH bits (except the UCI place- holder bits) are scrambled with pseudo-random sequence c(n) that is initialized with the 16- bit RNTI bits as follows 6(9) (i) = where Cinit = NRNTI215 + NID and nIDE{0,1, ., 1023} is set equal to the RRC parameter dataScramblingIdentityPUSCH oth- erwise, = Ncell The parameter NRNTI corresponds to the RNTI associated with the PUSCH transmission, if the input bits corresponding to UCI place-holder bits are set to one or the previous scrambled bit, depending on the value of the bits [6]. The block of scrambled bits are modulated using 2-BPSK (with transform precoding only), QPSK, 16QAM, 64QAM, or 256QAM modulation schemes, resulting in a block of complex-valued modulation symbol

s The complex-valued modulation symbols can be mapped to a maximum of four layers. More specifically, the complex-valued modulation symbols are mapped to x(i) = (i)...x(v-1)(i)] layer,i = 0, 1, where V is the number of layers and Nsymb layer is the number of modulation symbols per layer. If transform precoding is not enabled; for uplink CP-OFDM, y(1)(i) )=x(1)(i) for each layer l = 0, 1, V - 1. However, for DFT-S-OFDM uplink, V = 1, and ((i) depends on the configuration of phase-tracking ref- erence signals. If phase-tracking reference signals are not configured, the block of complex- valued symbols x(0)(0),. 1) for the single-layer v=1 are divided into Niaver INPUSCH sets, each corresponding to one OFDM symbol and = In case phase-tracking reference signals are configured, the block of complex-valued symbols (0) 1) are divided into a number of groups, where each group corre- sponds to one OFDM symbol. The 1th group contains NPUSCH subcarriers and is mapped to the complex-valued symbols (INPUSCH+i') corresponding to the 1th OFDM symbol prior to transform precoding, wherein i'= (0,1,..., NPUSCH The index m of PT-RS samples in the 1th group, the number of samples per PT-RS group Noroup , and the number of PT-RS groups NPT-RS are defined in [6]. The quantity E1 = 1, group when the 1th OFDM symbol contains one or more PT-RS samples, otherwise E1 = 0. New Radio Access Physical Layer Aspects (Part 2) 635 The transform precoding is then performed, resulting in a block of complex-valued symbols of the PUSCH in terms of resource blocks which must satisfy NPUSCH = 2a23a35as where A2, A3, a5 are non-negative integers. The DFT precoding is used to reduce the cubic metric of the uplink signal, thereby enabling higher power-amplifier efficiency. From implementa- tion point of view, it is better to constrain the DFT size to a power of 2. However, such a constraint would limit the scheduler flexibility in terms of the amount of resources that can be assigned for an uplink transmission. In NR, the DFT precoding size, and thus the

size of the resource allocation, is limited to products of the integers 2, 3, and 5 such that the DFT can be implemented as a combination of relatively less complex radix-2, radix-3, and radix- 5 FFT processing. The block of vectors Nsymb corresponding to layers precoded as follows: where = ap antenna port. For non- codebook-based transmission, the precoding matrix W is an identity matrix. However, for codebook-based transmission, the precoding matrix W is a scaler equal to one for single- layer transmission on a single-antenna port; otherwise, depending on value of the transmit- ted precoding matrix indicator (TPMI) index obtained from the DCI scheduling the uplink transmission, it will be chosen from a set of predefined matrices [6]. As an example, Table 4.32 provides the entries of the precoding vectors for single-layer transmission using two antenna ports. Table 4.32: Precoding matrix W for single-layer transmission using two antenna ports [6]. Index 0 Index 1 Index 2 Index 3 Index 4 Index 636 Chapter 4 For each antenna port used for transmission of PUSCH, the block of complex-valued sym- bols () (0), (p) (Maymb 1) are scaled by a factor of BPUSCH to adjust the transmit power and sequentially mapped, starting with () (0), to virtual resource elements (k', 1) in the vir- tual resource blocks assigned for PUSCH transmission. The physical resource blocks corre- sponding to the latter virtual resources must not be used for transmission of DM-RS, PT- RS, or DM-RS intended for other co-scheduled UEs. The mapping to virtual resource ele- ments (k',1) is in increasing order of frequency index k' over the assigned virtual resource blocks, where k'= 0 is the first subcarrier in the lowest numbered virtual resource block fol- lowed by time index 1. The virtual resource blocks are then mapped to physical resource blocks in a non-interleaved manner. For non-interleaved VRB-to-PRB mapping, virtual resource block n is mapped to physical resource block n. While dynamic scheduling is the basic mode of operation in NR, the res

ources for uplink data transmission or downlink data reception can be configured in advance for the UE. Once the uplink data are available at UE's buffer, it can immediately start uplink transmis- sion without going through the SR and grant cycle, thus reducing the latency. In other words, the NR PUSCH transmissions can be dynamically scheduled by an uplink grant pro- vided by a DCI, or the transmission can correspond to a configured grant Type 1 or Type 2. As shown in Fig. 4.88, the configured grant Type 1 PUSCH transmission is semi-statically configured to operate upon the reception of RRC parameter configuredGrantConfig includ- ing rrc-ConfiguredUplinkGrant without the detection of an uplink grant in a DCI. The con- figured grant Type 2 PUSCH transmission is semi-persistently scheduled by an uplink grant in a valid activation DCI after the reception of RRC parameter configurdGrantConfig that does not include rrc-ConfiguredUplinkGrant. The UE transmits PUSCH upon detection of a PDCCH with DCI format 0_0 or 0_1, and it is not expected to be scheduled to transmit RRC signaling: periodicity, offset, Possible uplink transmissions resources, activation Configured Activation periodicity Offset Configured grant type 1 RRC signaling: periodicity PDCCH: resources, activation Configured -Activation- periodicity Configured grant type 2 Figure 4.88 Illustration of uplink transmission with configured grants Type 1 and Type 2 [9]. New Radio Access Physical Layer Aspects (Part 637 another PUSCH by DCI format 0_0 or 0_1 scrambled by C-RNTI or MCS-C-RNTI for a given HARQ process until the end of the expected transmission of the last PUSCH for that HARQ process [9]. When the UE is scheduled to transmit a TB without a CSI report, or it is scheduled to trans- mit a TB with CSI report(s) on PUSCH by a DCI, the time domain resource assignment field value m of the DCI provides a row index m + 1 to an allocated table. The indexed row defines the slot offset K2, the start and length indicator SLIV, or directly the start symbol S and th

e allocation length L, and the PUSCH mapping type to be applied. Alternatively, when a UE is scheduled to transmit a CSI report(s) on PUSCH without a TB, the time domain resource assignment field value m of the DCI provides a row index m + to an allocated table which is defined by RRC parameter pusch-TimeDomainAllocationList in pusch-Config. The indexed row defines the start and length indicator SLIV, and the PUSCH mapping type. The parameter K2 value is determined as K2 = maxj Yj(m + 1), where Yj,j=0,.. Nrepetition - 1 are the corresponding list entries of the RRC parameter reportSlotOffsetList in CSI-ReportConfig for the Nrepetition triggered CSI Reporting Settings and Yj(m) is the mth entry of Yj [9]. The slot where the UE is expected to transmit PUSCH is determined by (2uPUSCH /2HPDCCH), + K2 where n is the slot with the scheduling DCI, K2 is based on PUSCH numerology, and UPUSCH and UPDCCH are the subcarrier spacing PUSCH and PDCCH, respectively. The starting symbol S relative to the start of the slot, and the number of consecutive symbols L counting from symbol S allocated for PUSCH are determined from the start and length indicator SLIV of the indexed row such that if (L - 1) <7 then SLIV = 14(L-1) + S; otherwise SLIV = 14(15 L) + (13 - S) where 14 - S. The PUSCH mapping type is set to Type A or Type B as given by the indexed row [9]. We have mentioned before that the time-domain reference point of the first PUSCH DM-RS symbol depends on the mapping type, where for PUSCH mapping type A, l is defined relative to the start of the slot, if frequency hopping is disabled and relative to the start of each hop in case frequency hopping is enabled, and the offset lo is given by the higher layer parameter dmrs-TypeA-Position. For PUSCH mapping type B, time-domain index l is defined relative to the start of the scheduled PUSCH resources, if frequency hop- ping is disabled and relative to the start of each hop when frequency hopping is enabled and the offset lo = 0 [6]. The UE determines the resource block assignment

in frequency domain using the resource allocation field in the detected PDCCH DCI except for Msg3 PUSCH initial transmission. Two uplink resource allocation types are supported. The uplink resource allocation Type 0 is supported for PUSCH only when transform precoding is disabled, whereas the uplink resource allocation Type 1 is supported for PUSCH regardless of whether transform precod- ing is disabled. If the scheduling DCI is configured to indicate the uplink resource alloca- tion type as part of the Frequency domain resource assignment field by setting RRC parameter resourceAllocation in pusch-Config to "dynamicswitch", the UE must use uplink 638 Chapter 4 resource allocation Type 0 or Type 1; otherwise, it uses the uplink frequency resource allo- cation type as defined by the RRC parameter resourceAllocation [9]. When the UE is scheduled with DCI format 0_0, the uplink resource allocation Type 1 is used. If a band- width part indicator field is not configured in the scheduling DCI, the RB indexing for uplink Type 0 and Type 1 resource allocation is determined within the UE's active band- width part. However, if a bandwidth part indicator field is configured in the scheduling DCI, the RB indexing for uplink Type 0 and Type 1 resource allocation is determined within the UE's bandwidth part indicated by bandwidth part indicator field value in the DCI. Upon detection of PDCCH, the UE first determines the uplink bandwidth part and then the resource allocation within the bandwidth part where RB numbering starts from the lowest RB in the determined uplink bandwidth part [9]. 4.2.5 Uplink MIMO Schemes The NR supports multi-antenna precoding with up to four layers for PUSCH transmission. However, when uplink DFT-based transform precoding is enabled, only single-layer (rank 1) transmission is supported. The UE can be configured in either codebook-based or non- codebook-based modes for PUSCH transmission. The selection between these two modes depends on the extent to which the uplink channel conditions can be estimate

d by the UE based on downlink measurements. PUSCH DM-RSs are precoded in the same way that PUSCH data subcarriers are precoded to allow coherent demodulation, making uplink pre- coding transparent to the gNB receiver. When codebook-based precoding is used in the uplink, the scheduling grant includes information about the precoder in the same way that the UE provides the network with PMI to assist downlink multi-antenna precoding. However, in contrast to the downlink, where the network may or may not use the precoder matrix indicated by the PMI, in the uplink direction, the UE is expected to use the precoder suggested by the network. In case of non-codebook-based transmission, the network can influence the selection of the uplink precoder. Another aspect that may impose constraints on the uplink multi-antenna transmission is to what extent one can assume coherence between different device antennas, that is, to what extent the relative phase between the sig- nals transmitted on two antennas can be controlled. The phase coherence is necessary when antenna port-specific weight factors, including specific phase shifts, are applied to the sig- nals transmitted on the different antenna ports. The NR specifications allow different UE capabilities concerning inter-antenna-port phase coherence, referred to as full coherence, partial coherence, and no coherence. In case of full coherence, it can be assumed that the device can control the relative phase between any of its antenna ports that are used for uplink transmission. In case of partial coherence, the device is capable of pairwise coher- ence, that is, the device can control the relative phase between antenna-port pairs. However, there is no guarantee that coherence can be achieved. In case of no coherence, there is no guarantee of phase coherence between any pair of the device antenna ports [6,9,14]. New Radio Access Physical Layer Aspects (Part 2) 639 In codebook-based uplink shared-channel transmission, the network selects the transmission rank and the corresponding

precoding matrix and informs the device through uplink sched- uling grant. At the UE side, the precoding matrix is applied to the scheduled PUSCH trans- mission and the indicated number of layers is mapped to the antenna ports. To select a suitable rank and a corresponding precoding matrix, the gNB needs estimates of the chan- nels between the device antenna ports and the corresponding network receive antennas. To enable this, a UE configured for codebook-based PUSCH transmission would typically be configured for transmission of at least one multi-port SRS. Based on measurements on the configured SRS, the network can estimate the channel and determine a suitable rank and precoding matrix. The network cannot select an arbitrary precoder, rather for a given combi- nation of the antenna ports and transmission rank, the network can select the precoding matrix from a limited set of available precoders [6,9,14]. When selecting the precoding matrix, the network needs to consider the device capability in terms of antenna-port phase coherence. If the UE does not support antenna-port phase coherence, only the first two precoding matrices can be used with rank-1 transmission. It must be noted that restricting the codebook selection to these two matrices is equivalent to selecting either the first or the second antenna port for transmission. In this type of antenna selection, phase coherence between the antenna ports is not required. The selec- tion of the remaining precoding vectors would imply linear combination of the signals of different antenna ports, which requires phase coherency among the antenna ports. A fun- damental difference between NR codebook-based PUSCH transmission and LTE uplink is that a device can be configured to transmit multiple SRS from multiple antenna ports. In multi-SRS transmission, the network feedback includes SRI, identifying one of the config- ured SRSs. The UE should then use the precoder identified in the scheduling grant and map the output of the precoder to the antenna ports corresponding

to the SRSs indicated in the SRI. The device should then transmit the precoded signal using the same antenna configuration and mapping that was used for the SRS transmission (indicated by the SRI). The use of multiple SRSs for codebook-based PUSCH transmission assumes that the UE transmits multi-port SRSs over separate and relatively wide beams. These beams may cor- respond to different UE antenna panels with different directions, where each panel includes a set of antenna elements corresponding to the antenna ports of each multi-port SRS (see Fig. 4.89). The SRI received from the network determines which beam should be used for the transmission, while the precoder information (number of layers and preco- der) determines how the transmission should be done within the selected beam. Codebook-based precoding is typically used when uplink/downlink reciprocity cannot be achieved and when uplink measurements are needed in order to determine a suitable uplink precoding [6,9,14]. In contrast to codebook-based precoding, which is based on network measurements and selection of uplink precoder, non-codebook-based precoding is based on device 640 Chapter 4 gNB Indicates to the UE the desired beam direction (SRS gNB Indicates to the UE the desired beam/precoder index), rank, and transmit precoding for the uplink direction and rank (all included in SRS indices) UE transmits multiple SRSs UE transmits multiple SRSs in different beam directions in different beam directions ((HII)) SRS3 / UE transmits as indicated by UE transmits uplink to match the direction of the the gNB indicated SRSs Codebook-based uplink transmission Non-codebook-based uplink transmission Figure 4.89 Codebook-based uplink transmission versus non-codebook-based uplink transmission [69]. measurements and precoder indications to the network. The concept of uplink non- codebook-based precoding is illustrated in Fig. 4.89. Based on downlink measurements con- ducted on configured CSI-RS resources, the UE selects a suitable uplink multi-layer preco- der. Non-cod

ebook-based precoding relies on channel reciprocity and assumes that the device can acquire accurate knowledge of the uplink channel based on downlink measure- ments. Note that there are no restrictions on the selection of precoder by the UE. Each col- umn of a precoding matrix defines a digital beam for the corresponding layer. Therefore, selection of precoder for each layer can be perceived as selection of different beam direc- tions, where each beam corresponds to one possible layer. It must be noted that the UE pre- coder selection is typically done based on downlink measurements, which may not be necessarily the best precoder from network point of view. As a result, the NR non- codebook-based precoding includes an additional step where the network can modify the device-selected precoder by removing some of the beams or equivalently some columns from the selected precoder [6,9,14]. As we mentioned earlier, codebook-based and non-codebook-based transmission modes are supported for PUSCH transmission. The UE will be configured with codebook-based trans- mission, when the RRC parameter txConfig in pusch-Config is set to "codebook", and it will be configured with non-codebook-based transmission, when the RRC parameter txConfig is set to "nonCodebook". If the RRC parameter txConfig is not provided, the PUSCH transmission will be based on a single-antenna port, triggered by DCI format 0_0. For codebook-based transmission, the UE determines the transmission precoder based on the information obtained from the SRI, TPMI, and the transmission rank, where the SRI, TPMI, and the transmission rank are given by the corresponding fields of the DCI. The TPMI is used to identify the preferred precoder over the SRS ports in the selected SRS resource by the SRI when single or multiple SRS resources are configured. Note that the indicated SRI in nth slot is associated with the most recent transmission of SRS resource identified by the SRI, where the SRS resource is prior to the PDCCH carrying the SRI. In codebook-based transmiss

ion mode, the UE determines its codebook subsets based on New Radio Access Physical Layer Aspects (Part 2) 641 TPMI and following reception of RRC parameter codebookSubset in pusch-Config, which may be configured with "fullyAndPartialAndNonCoherent" or "partialAndNonCoherent", or "nonCoherent" depending on the UE capability. The maximum transmission rank may be configured by the higher parameter maxRank in pusch-Config. Furthermore, for codebook- based transmissions, the UE may be configured with a single SRS-ResourceSet, and only one SRS resource can be indicated based on the SRI in the SRS resource set. The maximum number of configured SRS resources for codebook-based transmission is 2. If aperiodic SRS is configured for a UE, the SRS request field in DCI triggers the transmission of aperi- odic SRS resources [9]. The codebook subset restriction concept was introduced in LTE [15]. It helps avoid CSI reporting for the undesired (spatial) directions. The LTE codebook subset restriction includes RI and PMI restriction, which provides sufficient flexibility to control PMI calcula- tion and transmission from the UE. The content of CSI in NR is very similar to LTE, in the sense that NR supports CSI components such as RI and PMI. Since the number of possible RI values is small, bitmap with one-to-one correspondence between each bit in bitmap, and RI value can be specified for the purpose of RI restriction. At the same time, the number of possible PMI values especially for larger number of antenna ports is very large to support one-to-one correspondence between PMIs and bits in the bitmap. Therefore, a solution with reduced signaling overhead should be considered. More specifically, similar to LTE, a DFT beam restriction is introduced, SO that the PMI can be considered as restricted, if at least one beam is restricted by the corresponding DFT beam restriction bitmap. For co-phasing of the polarization, the bitmap should not be used as it does not affect the beamforming direction. The NR Type-I single-panel codebook str

ucture is similar to that of LTE FD-MIMO code- books, except rank 3/4 codebooks for 16, 24, and 32 antenna ports at the gNB. Let us con- sider codebook subset restriction for less than 16 antenna ports at the gNB. In this case, the beamforming vector for PMIs of all ranks is represented by 2D DFT beam denoted as bi, which is represented as Kronecker product of two 1D DFT vectors bi=unq wherein [66] e2nj((N,-1)/N101)n e27jj((N2-1)/N2O2)1 In this case, in order to restrict transmission in specific direction, a bitmap with size N1N2O1O2 can be specified, where each bit ai corresponds to DFT beam bi. If at least one layer of the PMI consists of bi, the PMI is considered to be restricted and cannot be 642 Chapter 4 Figure 4.90 Illustration of precoding for multi-panel antenna array [66]. reported by the UE. Rank 3/4 PMIs for 16, 24, and 32 antenna ports have different struc- tures compared to Type-I single-panel PMI. Therefore, it may not be possible to use the same approach for all ranks. The multi-panel codebook is constructed by DFT-based beam- forming per each panel and co-phasing of polarization and panels, where the same DFT beam is applied for all panels and polarizations. An example is shown in Fig. 4.90, where precoder p for rank-1 multi-panel codebook and two-panel antenna can be computed as fol- lows, in which C1, C2, C3 coefficients are independently reported in accordance to mode 2 of multi-panel codebooks [66]. In the preceding example, it is assumed that the antenna port indexing is performed in such way that [b c1b] corresponds to beamforming vector of the first polarization and [b c3/c2b] to beamforming vector of the second polarization. The direction of the transmission in the above PMI structure is determined by the DFT beam denoted by vector b and co-phasing coefficients C2 and C3/C2. Therefore, codebook subset restriction for multi-panel codebook PMI restriction consider all possible combina- tions of DFT beams and co-phasing coefficients, which are determining the direction of the transmission. T

he resulting size of bitmap in that case equals to [66]. The Type-II CSI was designed to enhance the performance of MU-MIMO transmission. The accuracy of spatial channel feedback in case of Type-II CSI allows interference sup- pression improvement through use of advanced precoding schemes such MMSE precoding. An accurate knowledge of the channel increases suppression capabilities of intra-cell inter- ference. The beamforming vector in Type-II codebook is represented as linear combination of 2, 3, or 4 DFT beams as Vki, where P(WB) Prli denotes the wideband beam amplitude scaling factor; Prli (SB) is the subband beam amplitude scaling factor, and Crli is the beam combining coefficient (phase) for beam i, polarization r, and layer l [66]. New Radio Access Physical Layer Aspects (Part 2) 643 For non-codebook-based uplink transmission, PUSCH transmission can be scheduled by DCI format 0_0, DCI format 0_1 or semi-statically configured. The UE can determine its PUSCH precoder and transmission rank based on the SRI when multiple SRS resources are configured, where the SRI is given by the SRI in DCI or the SRI is given by RRC parame- ter srs-ResourceIndicator. The UE must use one or multiple SRS resources for SRS trans- mission. In an SRS resource set, the maximum number of SRS resources which can be configured for the UE for simultaneous transmission on the same symbol and the maximum number of SRS resources depend on the UE capability. It must be noted that only one SRS port for each SRS resource can be configured and only one SRS resource set can be config- ured with higher layer parameter usage in SRS-ResourceSet set to "nonCodebook". The maximum number of SRS resources that can be configured for non-codebook-based uplink transmission is 4. The indicated SRI in the nth slot is associated with the most recent trans- mission of SRS resource(s) identified by the SRI, where the SRS transmission is prior to the PDCCH carrying the SRI [9]. For non-codebook-based uplink transmission, the UE can calculate the precoder used f

or the transmission of SRS based on measurement of an associated NZP CSI-RS resource. A UE can be configured with only one NZP CSI-RS resource for the SRS resource set with higher layer parameter usage in SRS-ResourceSet set to "nonCodebook". If aperiodic SRS resource set is configured, the associated NZP-CSI-RS is indicated via SRS request field in DCI formats 0_1 and 1_1. A UE is not expected to update the SRS precoding information, if the gap between the last symbol of the reception of the aperiodic NZP-CSI-RS resource and the first symbol of the aperiodic SRS transmission is less than 42 OFDM symbols. The CSI-RS is located in the same slot as the SRS request field. If the UE is configured with aperiodic SRS associated with aperiodic NZP CSI-RS resource, none of the TCI states con- figured on the scheduled component carrier are configured with "QCL-TypeD". The UE performs one-to-one mapping between the indicated SRI(s) and the indicated DM-RS port (s) and their corresponding PUSCH layers provided by DCI format 0_1 or by configuredGrantConfig. The UE transmits PUSCH using the same antenna ports as the SRS port(s) in the SRS resource(s) indicated by SRI(s) given by DCI format 0_1 or by configuredGrantConfig. For non-codebook-based uplink transmission, the UE can be sched- uled with DCI format 0_1 when at least one SRS resource is configured in SRS-ResourceSet with usage set to "nonCodebook" [9]. 4.2.6 Link Adaptation and Power Control Power control is a mechanism where the transmit power of the downlink or uplink control or traffic channels are adjusted at the gNB or at UEs, based on instructions from the serving base station such that with minimal impact on the reliability of the downlink/uplink trans- missions and throughput, the inter-user/inter-cell interference among users and base stations 644 Chapter 4 are reduced. Therefore, power control can be considered as a link adaptation mechanism that is utilized for interference mitigation in cellular systems. While increasing the transmit power over a communicat

ion link has certain advantages such as higher SNR at the receiver, which reduces the BER and allows higher data rate and results in greater spectral efficiency as well as more protection against signal attenuation over fading channels, a higher transmit power; however, has several drawbacks including increased power consumption of the transmitting device, reducing the UE battery life, and increased interference to other users in the same or adjacent frequency bands. The following sections describe the power control algorithms that are incorporated in NR. The NR provides uplink power control mechanisms to compensate the effects of path loss, shadowing, fast fading, and implementation loss. The uplink power control is further used to mitigate inter-cell and intra-cell interference, thereby enhancing the overall throughput and reducing the effective UE power consumption. The uplink power control includes open- loop and closed-loop power control. The base station transmits necessary power control information through transmission of power control messages. The parameters of the power control algorithm are optimized on a system-wide basis by the gNB and are broadcast peri- odically or trigged by events. The UE provides the necessary information through higher layer control messages to the serving gNB in order to enable uplink power control. The gNB can exchange necessary information with neighboring base stations through backhaul to support uplink power control to facilitate the handover process. The power control scheme may not be effective in high mobility scenarios for compensating the effects of a fast fading channel due to variation of the channel impulse response. As a result, the power control is used to mitigate the distance-dependent path loss, shadowing, and implementation loss. The uplink power control takes into consideration the MIMO transmission mode and whether a single user or multiple users are supported on the same resource at the same time. The open-loop power control compensates the channel variati

ons and implementation loss without requiring frequent interactions with the serving gNB. The UE can determine the transmit power based on the transmission parameters sent by the gNB, uplink channel quality, downlink CSI, or the interference knowledge obtained from downlink transmissions. The open-loop power control provides a coarse initial transmit power setting for the device before establishing connection with the base station. It is believed that rate control is more efficient than power control under certain conditions. Rate control in principle implies that the power amplifier is always transmitting at full power and therefore is efficiently utilized. On the other hand, power control often results in inefficient utilization of the power amplifier because the transmission power is often less than its maxi- mum. In practice, the radio-link data rate is controlled by adjusting the modulation scheme and/or the channel coding rate. In good channel conditions, the value of Eb/No at the receiver is high, and the main limitation of the data rate is the bandwidth of the radio link. New Radio Access Physical Layer Aspects (Part 2) 645 In such conditions, use of higher order modulation, for example, 16QAM 64QAM, or 256QAM together with a high coding rate, is more appropriate for link adaptation. Similarly, in the case of poor channel conditions, the use of QPSK and low-rate coding is preferred. Link adaptation by means of rate control is referred to as adaptive modulation and coding. A power control mechanism takes into consideration the serving gNB target link SINR and/ or interference level to other cells/sectors for mitigating inter-cell interference. In order to achieve the target SINR, the serving gNB path loss can be fully or partially compensated based on a trade-off between overall system throughput and cell-edge performance. The UE transmit power is adjusted in order to ensure the level of interference is less than the per- missible interference level. The closed-loop power control, on the other hand, compen

sates channel variations through periodic power-control commands from the serving gNB. The base station measures the uplink CSI and interference level using uplink data and/or control channel transmissions and sends power control commands to the devices. Upon receiving the power control command from the gNB, the UE adjusts its uplink transmit power. The closed-loop power control is active during data and control channel transmissions. A UE is expected to maintain the transmit power density (i.e., total transmit power normalized by transmission bandwidth) for each data and control channel below a certain level that is determined by the maximum permissible power level for the UE, emission mask, and other regulatory constraints. In other words, when the number of active logical resource units assigned to a particular user is reduced, the total transmitted power must be reduced propor- tionally by the UE in the absence of any additional change of power control parameters. When the number of resource blocks is increased, the total transmitted power must be pro- portionally increased such that the transmitted power level does not exceed the permissible power levels specified by 3GPP and the regulatory specifications [1-3]. For interference level control, the information about the current interference level of each gNB may be shared among the base stations via backhaul. The (uplink) TPC in mobile communication systems is meant to balance the transmitted energy per bit in order to maintain the link quality corresponding to the minimum QoS requirements, to minimize interference to other users in the system, and to minimize the power consumption of the device. In achieving these goals, the power control has to adapt to the characteristics of the propagation channel, including path loss, shadowing, and fast fading, as well as overcoming interference from other users both within the same cell and in neighboring cells. The NR uplink power control is similar to LTE and is based on a com- bination of open-loop power control, in

cluding support for fractional path loss compensa- tion, where the device estimates the uplink path loss based on downlink measurements and sets the transmit power accordingly, and closed-loop power control based on explicit power-control commands provided by the network. In practice, these power-control 646 Chapter 4 commands are determined based on prior measurements of the received uplink power. The main difference with the LTE control is the possibility of beam-based power control. Uplink power control determines the power level for PUSCH, PUCCH, SRS, and PRACH transmissions. The ith PUSCH/PUCCH/SRS/PRACH transmission occasion is defined by slot index Nslot within a frame with system frame number SFN, the first symbol S within the slot, and the number of consecutive symbols L. For a PUSCH transmission on active uplink BWP k of carrier l of serving cell m and parameter set configuration with index j and PUSCH power control adjustment state with index , a UE first calculates a linear value of the transmit power PUSCH(i,j,9d,u). If PUSCH transmission is sched- uled by DCI format 0_1 and when txConfig in PUSCH-Config is set to "codebook", the UE scales the linear value by the ratio of the number of antenna ports with a non-zero PUSCH transmission power to the maximum number of SRS ports supported by the UE in one SRS resource. The UE divides the power equally among the antenna ports on which it transmits PUSCH with non-zero power. The PUSCH transmit power is calculated as follows [8]: where (i,j, qd, denote transmission occasion, parameter set configuration index, reference signal index for the active downlink BWP, and PUSCH power control adjustment state index, respectively. PCMAXIm (i) denotes the maximum permissible UE transmit power for carrier 1, serving cell m, and in PUSCH transmission occasion i. Po_PUSCH(j) = P0_NOMINAL_PUSCH,MU + Po_UE_PUSCHgim je {0, J - 1} is parameter determined by RRC parameters preambleReceivedTargetPower, msg3- DeltaPreamble, ConfiguredGrantConfig, pO-NominalWithGrant p0-PUSCH-Alp

ha, PO- PUSCH-AlphaSet, SRI-PUSCH-PowerControl as well as the SRI field in DCI format 0_0 and 0_1. The quantity Po is provided as part of the power-control configuration and would typically depend on the target data rate but also on the noise and interference level experienced at the receiver. NPUSCH(i) denotes the bandwidth of PUSCH resource assignment expressed in number of resource blocks. akim(j) is a network-configurable parameter corresponding to fractional path loss compensation which is determined by RRC parameters msg3-Alpha, ConfiguredGrantConfig, p0-PUSCH-Alpha, PO-PUSCH-AlphaSet, SRI-PUSCH- PowerControl as well as the SRI field in DCI format 0_0 and 0_1. In the case of frac- tional path loss compensation (a <1), the path loss will not be fully compensated, and New Radio Access Physical Layer Aspects (Part 2) 647 the average received power will vary depending on the location of the device within the cell. In this case, the received power is lower for devices with higher path loss located at larger distances from the cell site. This must then be compensated by adjusting the uplink data rate. The advantage of fractional path loss compensation is reduced interfer- ence to neighboring cells, which is achieved at the expense of larger variation in the ser- vice quality, with reduced peak data-rate availability for devices closer to the cell edge. PLklm(qd) is the downlink path loss estimate in dB calculated by the UE using reference signal index qd for the active downlink BWP. for Ks = 1.25 (provided by deltaMCS); oth- erwise related to the modulation scheme and channel coding rate used for the PUSCH transmission. The parameter bits per resource element (BPRE) is given by BPRE = -1 Kr/NRE for PUSCH with UL-SCH data and BPRE = QmR/Bottself offset for CSI transmission in PUSCH without UL-SCH data wherein Qm is the modulation order and R is the target code rate. This term models how the required received power varies when the number of information BPRE changes due to different modulation schemes and channel-co

ding rates. fklm(i,u) denotes PUSCH power control adjustment state which is given by &PUSCH(.) is the power adjustment due to closed-loop power control. The power control commands can be sent to multiple devices by means of DCI format 2_2. Each power control command consists of 2 bits corresponding to four different update steps (-1, 0, +1, +3 dB). The reason for including 0 dB as an update step is that a power-control command is included in every scheduling grant, and it is desirable not to have to adjust the PUSCH transmit power for each grant. The PUCCH power control follows the same principles as PUSCH power control with some minor differences. For PUCCH power control, there is no fractional path loss compensation (a = 1). Furthermore, for PUCCH power control, the closed-loop power control commands are carried within DCI formats 1_0 and 1_1, which are used for downlink scheduling assignments rather than within uplink scheduling grants. This is partly due to the fact that PUCCH transmission is used to carry HARQ-ACKs in response to a downlink transmission and such downlink transmissions are typically associated with downlink scheduling assignments on PDCCH and the corresponding power control commands could be used to adjust the PUCCH transmit power prior to the transmission of HARQ-ACKs. Similar to PUSCH, power control commands can also be carried jointly to multiple devices by means of DCI format 2_2. When the UE transmits PUCCH on active uplink BWP k of carrier l of serving cell m and PUCCH power control adjustment state with index U, it determines the PUCCH transmission power in the ith PUCCH transmission occasion as follows [8]: 648 Chapter 4 where PCMAXIm (i) is the configured UE transmit power. Po_PUCCHm(qu) = Po_NOMINAL_PUCCH + Po_UE_PUCCH(Qu) whose parameters are pro- vided p0-nominal and p0-PUCCH-Value. NPUCCH( RBklm is the bandwidth of PUCCH resource assignment expressed in number of resource blocks. PLklm(qd) is the downlink path loss estimate in dB calculated by the UE using reference signal resour

ce index qd. AF_PUCCH(F) is PUCCH format-dependent power adjustment parameter provided by deltaF-PUCCH-f0 for PUCCH format 0, deltaF-PUCCH-fI for PUCCH format 1, deltaF-PUCCH-f2 for PUCCH format 2, deltaF-PUCCH-f3 for PUCCH format 3, and deltaF-PUCCH-f4 for PUCCH format 4. \TFklm(i) is the PUCCH transmission power adjustment component which is dependent on the PUCCH format and the corresponding transport parameters. gklm(i,u) denotes the PUCCH power control adjustment state. The UE calculates the transmission power for PRACH as PPRACHkm (i) = min(PCMAX, (i), = P PRACH-Targetim + PLklm) where PCMAX (i) is the configured maximum UE transmission power, P PRACH-Target is the target PRACH reception power corresponding to PREAMBLE_RECEIVED_TARGET_POWER parameter provided via RRC signaling, and PLklm is path loss estimated for the active uplink BWP k of carrier l of serving cell m. The path loss is estimated based on the downlink reference signal associated with the PRACH transmission on the active downlink BWP of the mth serving cell and it is calculated by the UE as [reference signal power] - [higher layer filtered RSRP] in (dBm). If the active downlink BWP is the initial downlink BWP and for the SS/PBCH block and CORESET multiplexing pattern 2 or 3, the UE determines PLklm based on the SS/PBCH block associ- ated with the PRACH transmission [8]. The UE can set its configured maximum output power PCMAX|m for the 1th carrier of the mth serving cell in each slot. The configured maximum output power PCMAXIm is set within PowerClass - max (MPRm + A-MPRm P-MPRM)] and PCMAX-Highlm = in(PEMAX PpowerClass - \PPowerClass). In the latter equa- tion, PEMAX,, is the value given by information element P-Max for the mth serving cell, and PpowerClass is the maximum UE power without taking into account the tolerance specified by 3GPP specifications [1,2]. The PRACH preamble transmission involves some uncertainty about the required transmit power. As a result, the PRACH preamble transmission includes a power-ramping New Radio Access Ph

ysical Layer Aspects (Part 2) 649 mechanism where the preamble may be retransmitted with a transmit power that is increased in steps during each transmission attempt. The device selects the initial PRACH preamble transmit power based on estimates of the downlink path loss in combination with a target received preamble power configured by the network. The path loss should be esti- mated based on the received power of the SS/PBCH block that the device has acquired and has determined the RACH resources for preamble transmission. If no random-access response is received within a predetermined window, the device can assume that the pream- ble was not correctly received by the network. In that case, the device repeats the preamble transmission with an increased transmit power. This power ramping continues until a random-access response has been received or until a configurable maximum number of retransmissions are attempted. Under such condition the random-access transmission has failed [14]. When uplink beamforming is used, the uplink path loss estimate PLklm(qd) is used to deter- mine the transmit power. The latter path loss estimate includes the effect of beamforming gain of the uplink beam pair to be used for PUSCH transmission. Assuming beam corre- spondence, this can be achieved by estimating the path loss based on measurements on a downlink reference signal transmitted over the corresponding downlink beam pair. As the uplink beam used for the transmission pair may change between PUSCH transmissions, the device may have to perform multiple path loss estimates corresponding to different candi- date beam pairs. During PUSCH transmission over a specific beam pair, the path loss esti- mate corresponding to that beam pair is used to determine PUSCH transmit power. This is enabled by the parameter q in the path loss estimate PLkim(qd)[14]. The gNB configures the UE with a set of downlink reference signals based on which the path loss is estimated. Each reference signal is associated with a specific value of q. To limit

the number of path loss estimations, the UE is not required to perform more than four parallel path loss estimations corresponding to different directions. The network configures a mapping between possible SRI values, provided in the scheduling grant, and different values of q. When a PUSCH transmission is scheduled by a scheduling grant including SRI, the path loss estimate associated with that SRI is used for determining the transmit power for the scheduled PUSCH transmission [14]. In the preceding equation for PUSCH power control, the open-loop parameters Po_PUSCHkim (j) and aklm(j) are associated with parameter j, suggesting that there are multiple open-loop parameter pairs that can be used for different types of PUSCH transmissions, for example, Msg3 on PUSCH, grant-free PUSCH transmission, and scheduled PUSCH trans- mission. However, there is also a possibility to have multiple pairs of open-loop parameter for scheduled PUSCH transmission, where the pair to use for a certain PUSCH transmission can be selected based on the SRI similar to the selection of path loss estimates. In practice, this implies that the open-loop parameters Po_PUSCHm and akim(j) will depend on the uplink beam [14]. 650 Chapter 4 For PUSCH transmissions, the device can be configured with different open-loop parameter pairs Po_PUSCHkm (j) and akim(j) corresponding to different values of parameter j, where j = 0 is associated with Msg3 transmission and j = 1 is used in the case of grant-free PUSCH transmission. Each possible value of the SRI that can be provided as part of the uplink scheduling grant is associated with one of the configured open-loop parameter pairs. When a PUSCH transmission is scheduled with a certain SRI included in the scheduling grant, the open-loop parameters associated with that SRI are used when determining the transmit power for the scheduled PUSCH transmission [8]. The other parameter in PUSCH power control equation is the power control adjustment state with index which is related to the closed-loop mechanism.

PUSCH power control allows two independent closed-loop processes associated with U = 0 and = 1. The value(s) are provided by sri-PUSCH-ClosedLoopIndex RRC parameter. If PUSCH transmission is scheduled by a DCI format 0_1 and if DCI format 0_1 includes an SRI field, the UE determines the value that is mapped to the SRI field. This means that similar to the case for multiple path loss estimates and multiple open-loop parameter pairs, the selec- tion of indicates the selection of the closed-loop process which is associated to the SRI included in the scheduling grant [8]. As we mentioned earlier, the UE relies on downlink measurements to calculate the path loss and to determine the power control parameters. The gNB determines the downlink transmit EPRE. For SS-RSRP, SS-RSRQ, and SS-SINR measurements, the UE may assume that downlink EPRE is constant across the bandwidth. The UE may further assume that down- link EPRE is constant over SSS carried in different SS/PBCH blocks, and the ratio of SSS EPRE to PBCH DM-RS EPRE is 0 dB [8]. For CSI-RSRP, CSI-RSRQ, and CSI-SINR mea- surements, the UE may assume downlink EPRE of a port of CSI-RS resource configuration is constant across the configured downlink bandwidth and constant across all configured OFDM symbols. The downlink SS/PBCH SSS EPRE can be derived from the SS/PBCH downlink transmit power given by the parameter SS-PBCH-BlockPower provided by RRC signaling. The downlink SSS transmit power is defined as the linear average over the power contributions (in Watts) of all REs that carry the SSS within the operating system band- width. The downlink CSI-RS EPRE can be derived from the SS/PBCH block downlink transmit power given by the parameter SS-PBCH-BlockPower and CSI-RS power offset given by the parameter powerControlOffsetSS provided through RRC signaling. The down- link reference signal transmit power is defined as the linear average over the power contri- butions [in (W)] of the resource elements that carry the configured CSI-RS within the operating system bandwidth.

For the purpose of downlink power allocation, the ratio of PDSCH EPRE to DM-RS EPRE (PDM-RS dB]) is given in Table 4.33 for downlink DM-RS associated with PDSCH, which depends on the number of DM-RS CDM groups without data [9]. It can be shown that the New Radio Access Physical Layer Aspects (Part 2) 651 Table 4.33: Ratio of PDSCH EPRE to DM-RS EPRE [9]. Number of DM-RS CDM DM-RS Configuration DM-RS Configuration Groups without Data Type 1 (dB) Type 2 (dB) - 4.77 Table 4.34: PT-RS EPRE to PDSCH EPRE per layer per resource element [9]. EPRE-Ratio PPT-RS Number of PDSCH Layers Reserved Reserved DM-RS power scaling factor applied prior to DM-RS resource mapping is given by BDM-RS PDSCH = 10(-PDM-Rs/20). When the UE is scheduled with a PT-RS port associated with the PDSCH, the ratio of PT- RS EPRE to PDSCH EPRE per layer per resource element for PT-RS port PPT-RS is given by Table 4.34. In that case, the PT-RS power scaling factor BPT-RS is given by BPT-RS = 10 PPT-RS/20); otherwise, it can be assumed that epre-Ratio is set to state "0" in Table 4.34. References 3GPP Specifications [1] 3GPP TS 38.101-1, NR, User Equipment (UE) Radio Transmission and Reception; Part 1: Range 1 Standalone (Release 15), December 2018. [2] 3GPP TS 38.101-2: NR, User Equipment (UE) Radio Transmission and Reception; Part 2: Range 2 Standalone (Release 15), December 2018. 3GPP TS 38.104, NR, Base Station (BS) Radio Transmission and Reception (Release 15), December 2018. 3GPP TS 38.133, NR, Requirements for Support of Radio Resource Management (Release 15), December 2018. 3GPP TS 38.202, NR, Services Provided by the Physical Layer (Release 15), December 2018. [6] 3GPP TS 38.211, NR, Physical Channels and Modulation (Release 15), December 2018. 3GPP specifications can be accessed at the following URL: http://www.3gpp.org/ftp/Specs/archive/. 652 Chapter 4 [7] 3GPP TS 38.212, NR, Multiplexing and Channel Coding (Release 15), December 2018. [8] 3GPP TS 38.213, NR, Physical Layer Procedures for Control (Release 15), December 2018. [9] 3GPP TS 38

.214, NR, Physical Layer Procedures for Data (Release 15), December 2018. [10] 3GPP TS 38.215, NR, Physical Layer Measurements (Release 15), December 2018. [11] 3GPP TS 38.300, NR, NR and NG-RAN Overall Description, Stage 2 (Release 15), December 2018. [12] 3GPP TS 38.321, NR, Medium Access Control (MAC) Protocol Specification (Release 15), December 2018. [13] 3GPP TS 38.331, NR, Radio Resource Control (RRC), Protocol Specification (Release 15), December 2018. Articles, Books, White Papers, and Application Notes [14] E. Dahlman, S. Parkvall, 5G NR: The Next Generation Wireless Access Technology, Academic Press, August 2018. [15] S. Ahmadi, LTE-Advanced: A Practical Systems Approach to Understanding 3GPP LTE Releases 10 and 11 Radio Access Technologies, Academic Press, November 2013. [16] T.L. Marzetta, et al., Fundamentals of Massive MIMO, Cambridge University Press, December 2016. [17] T.S. Rappaport, R.W. Heath Jr, Millimeter Wave Wireless Communications, Prentice Hall, September, 2014. S. Lin, D.J. Costello, Error Control Coding, second ed., Prentice Hall, June 2004. [19] C.E. Shannon, W. Weaver, The Mathematical Theory of Communication, University of Illinois Press, September 1949. [20] A. Papoulis, Probability, Random Variables and Stochastic Processes, fourth ed., McGraw Hill Higher Education, January 2002. [21] T.M. Cover, J.A. Thomas, Elements of Information Theory, John Wiley & Sons, July 2006. [22] J.G. Proakis, Digital Communications, fifth ed., McGraw-Hill, November 2007. [23] F. Ademaj, et al., 3GPP 3D MIMO channel model: a holistic implementation guideline for open source simulation tools, EURASIP J. Wirel. Commun. Netw. 2016 (2016) 55. [24] R.G. Gallager, Principles of Digital Communication, Cambridge University Press, March 2008. [25] B. Mondal, et al., 3D channel model in 3GPP, IEEE Commun. Mag. 53 (3) (2015). [26] F. Vook, 3GPP new radio at sub-6GHz: features and performance, in: IWPC Workshop on 5G NR Mobile Networks and User Equipment, January 2018. [27] T.S. Rappaport, et al., Overview of mil

limeter wave communications for fifth-generation (5G) wireless networks-with a focus on propagation models, in: IEEE Transactions on Antennas and Propagation, Special Issue on 5G, November 2017. [28] S. Cammerer, Spatially Coupled LDPC Codes, Institut für Nachrichtenübertragung. WebDemos. Available from: <http://www.inue.uni-stuttgart.de/lehre/demo.html> [29] A. Elkelesh, M. Ebada, Polar Codes, Institut für Nachrichtenübertragung. WebDemos. Available from: <http://www.inue.uni-stuttgart.de/lehre/demo.html>. [30] 5G New Radio, ShareTechNote. Available from: <http://www.sharetechnote.com) [31] J. Campos, Understanding the 5G NR Physical Layer, Keysight Technologies, November 2017. MediaTek, 5G NR: A New Era for Enhanced Mobile Broadband, White paper, March 2018. [33] E. Arikan, Channel polarization: a method for constructing capacity achieving codes for symmetric binary-input memoryless channels, IEEE Trans. Inf. Theory 55 (7) (2009) 3051-3073. E. Arikan, Systematic polar coding, IEEE Commun. Lett. 15 (8) (2011) 860-862. [35] R.G. Maunder, The implementation challenges of polar codes, in: AccelerComm White Paper, February 2018. [36] V. Bioglio, C. Condo, I. Land, Design of Polar Codes in 5G New Radio, Cornell University Library, 2019. [37] F. Hamidi-Sepehr, et al., Analysis of 5G LDPC codes rate-matching design, in: IEEE Vehicular Technology Conference (VTC), June 2018. New Radio Access Physical Layer Aspects (Part 2) [38] 3GPP TSG RAN WG1, R1-1608862, Polar Code Construction for NR, Huawei, HiSilicon, 2016. [39] R.G. Gallager, MIT Press Classic Series Low-Density Parity-Check Codes, MIT Press, Cambridge, MA, 1963. [40] F. Sabatier, Polar Coding Tutorial. Available from: <http://ipgdemos.epfl.ch/polarcodestutorial/index. html>. [41] C. Berrou, A. Glavieux, P. Thitimajshima, Near Shannon limit error-correcting coding and decoding: turbo-codes, in: Proceedings of IEEE International Communications Conference (ICC), May 1993. [42] M. Giordani, et al., A tutorial on beam management for 3GPP NR at mmWave frequencies, IEE

E Commun. Surv. Tutorials (2018). [43] H. Shariatmadari, et al., Fifth-generation control channel design, achieving ultra-reliable low-latency com- munications, IEEE Veh. Technol. Mag. 13 (2) (2018) 84-93. 5G New Radio: Introduction to the Physical Layer, White Paper, National Instruments, 2018. [45] S.M. Razavizadeh, et al., Three-dimensional beamforming: a new enabling technology for 5G wireless networks, IEEE Signal Process Mag. 31 (6) (2014) 94-101. H. Ji, et al., Overview of full-dimension MIMO in LTE-Advanced Pro, IEEE Commun. Mag. 55 (2) (2017) 176-184. [47] Y.-H. Nam, et al., Full dimension MIMO for LTE-advanced and 5G, in: Information Theory and Applications Workshop (ITA), February 2015. [48] A. Alkhateeb, G. Leus, R.W. Heath, Multi-layer precoding: a potential solution for full-dimensional mas- sive MIMO systems, IEEE Trans. Wireless Commun. 16 (9) (2017) 5810-5824. J. Jeon, NR wide bandwidth operations, IEEE Commun. Mag. 56 (3) (2018) 42-46. [50] D.H.N. Nguyen, T. Le-Ngoc, MMSE precoding for multiuser MISO downlink transmission with non-ho- mogeneous user SNR conditions, EURASIP J. Adv. Signal Process. 2014 (2014) 85. [51] R.-A. Pitaval, et al., Overcoming 5G PRACH capacity shortfall by combining Zadoff-Chu and M-sequences, IEEE Int. Conf. Commun. (ICC) (2018). [52] T.L. Marzetta, Massive MIMO: an introduction, Bell Labs Tech. J. 20 (2015) 11-22. [53] F. Hamidi-Sepehr, et al., 5G NR PDCCH: design and performance, in: 2018 IEEE 5G World Forum (5GWF), July 2018. [54] Y. Qi, et al., On the Phase Tracking Reference Signal (PT-RS) Design for 5G New Radio (NR), Cornell University Library, 2018. Y. Blankenship, et al., Channel coding in 5G new radio: a tutorial overview and performance comparison with 4G LTE, IEEE Veh. Technol. Mag. 13 (4) (2018) 60-69. [56] M. Reil, G. Lloyd, Millimeter-wave beamforming: antenna arrays & characterization, White Paper, Rohde & Schwarz, 2016. E. Onggosanusi, et al., Modular and high-resolution channel state information and beam management for 5G new radio, IEEE Commun. Mag. 5

6 (3) (2018) 48-55. [58] A.L. Swindlehurst, et al., Millimeter-wave massive MIMO: the next wireless revolution? IEEE Commun. Mag. 52 (9) (2014) 56-62. [59] Y. Huang, et al., Multi-panel MIMO in 5G, IEEE Commun. Mag. 56 (3) (2018) 56-61. [60] T.L. Marzetta, Noncooperative cellular wireless with unlimited numbers of base station antennas, IEEE Trans. Wireless Commun. 9 (11) (2010) 3590-3600. [61] F. Rusek, et al., Scaling up MIMO: opportunities and challenges with very large arrays, IEEE Signal Process Mag. 30 (1) (2013) 40-60. [62] E.G. Larsson, et al., Massive MIMO for next generation wireless systems, IEEE Commun. Mag. 52 (2) (2014) 186-195. [63] L. Lu, et al., An overview of massive MIMO: benefits and challenges, IEEE J. Sel. Top. Signal Process. 8 (5) (2014) 742-758. W. Roh, et al., Millimeter-wave beamforming as an enabling technology for 5G cellular communications: theoretical feasibility and prototype results, IEEE Commun. Mag. 52 (2) (2014) 106-113. 654 Chapter 4 [65] 3GPP TSG RAN WG1, R1-1800036, Summary of Remaining Issues on HARQ Management, Huawei, HiSilicon, 2018. [66] 3GPP TSG RAN WG1, R1-1712546, Discussion on Codebook Subset Restriction for NR, Intel Corporation, 2017. [67] 5G Toolbox, The MathWorks, Inc. Available from: <https://www.mathworks.com/help/5g/index.html> [68] H. Koorapaty, NR physical layer design: physical layer structure, numerology and frame structure, in: RWS-180007, Workshop on 3GPP Submission Towards IMT-2020, October 2018. [69] Y. Kim, NR physical layer design: NR MIMO, in: RWS-180008, Workshop on 3GPP Submission Towards IMT-2020, October 2018. [70] G. Noh, et al., DM-RS design and evaluation for 3GPP 5G new radio in a high speed train scenario, IEEE Global Commun. Conf, 2017. X. Lin, et al., 5G New Radio: Unveiling the Essentials of the Next Generation Wireless Access Technology, Cornell University Library, 2018. [72] J. Liu, et al., Initial access, mobility, and user-centric multi-beam operation in 5G new radio, IEEE Commun. Mag. 56 (3) (2018) 35-41. Updated information on Chi

[74] W.U. Yong, Self evaluation: enhanced mobile broadband (eMBB) evaluation results, in: RWS-180018, Workshop on 3GPP Submission Towards IMT-2020, October 2018. [75] D.J. Love, R.W. Heath Jr., Grassmannian precoding for spatial multiplexing systems, in: Proceedings of the Allerton Conference on Communication Control and Computing, Monticello, 2003. [76] N. Jindal, MIMO broadcast channels with finite-rate feedback, IEEE Trans. Inf. Theory 52 (11) (2006). CHAPTER 5 New Radio Access RF and Transceiver Design Considerations The NR RF transceiver characteristics are related to the frequency bands in which the 5G systems will be deployed. Due to wide range of the target frequency bands, spectrum flexibility was required for the new radio in order to operate in diverse spectrum alloca- tions. While spectrum flexibility has been used in the previous generations of radio access technologies, it has become more important for NR development and deployment. Such spectrum flexibility is manifested as feasibility of deployment and resource alloca- tions in frequency blocks of different sizes over an extremely wide range of contiguous or non-contiguous spectrum, both in the form of paired and unpaired frequency bands along with aggregation of different spectrum blocks within and across different bands. The NR has the capability to operate with mixed OFDM numerologies over the same or different RF carrier(s) and has relatively more flexibility compared to LTE in terms of frequency-domain scheduling and multiplexing of devices over the serving base station (BS) RF carrier(s). The use of OFDM waveform in NR provides the desired flexibility in terms of the size of the spectrum allocation and the instantaneous transmission bandwidth adaptation. The application of active antenna system (AAS) concept and multiple anten- nas in the base stations and the devices, which emerged during LTE development, has taken a giant leap in NR with the support of massive MIMO and control/data channel beamforming both in the existing LTE bands and in

the new mmWave bands. Aside from physical layer design implications, the advent of the latter features significantly impact the analog/digital RF hardware system design/implementation including filters, ampli- fiers, data converters, antennas, etc. In this chapter, we will discuss the new radio spectrum, RF characteristics, implementation considerations of the NR base stations and the devices as well as the hardware technologies that are used to implement various features of the new radio. In the course of this chapter, we will discuss various types of NR base stations and their external interfaces over which the RF requirements are defined. We further explore the conducted and over-the-air (OTA) RF requirements specified by 3GPP for testing and evaluating the performance of the NR base stations and devices. 5G NR. DOI: https://doi.org/10.1016/B978-0-08-102267-2.00005-1 © 2019 Elsevier Inc. All rights reserved. 656 Chapter 5 5.1 NR Radio Parameters and Spectrum The requirement for spectrum flexibility was a key driving factor for the adoption of OFDM-based technologies in 3GPP LTE which continues to be a major driver for the NR frequency planning and deployments. The need for diverse spectrum allocations in terms of spectral bands, operation bandwidths, duplex schemes (paired and unpaired spectrum) and multiple-access schemes emerged during the 3G and 4G deployments and has led to one of the most distinctive characteristics of 5G NR, which supports a large and diverse spectrum from 450 MHz to 52.6 GHz (and up to 100 GHz in future releases). The maximum fre- quency currently under study in ITU-R is 86 GHz (see Fig. 5.1). The NR supports operating bandwidths of 5 MHz to 3.2 GHz for both paired and unpaired spectrum as well as supple- mentary downlink (SDL) or supplementary uplink (SUL) carriers. The NR defines various frequency bands within the 5G spectrum. Although the boundaries of the NR frequency bands can vary in different countries and regions, it must be possible to efficiently allocate RF carriers at positi

ons where the spectrum blocks are used with minimal spectrum wastage. This requires carefully defined channel raster for carrier allocation. There are also a number of other spectrum blocks being considered by regulatory bodies that were not initially con- sidered by 3GPP specifications, which include 5925-7150 MHz in the United States and 5925-6425 MHz by CEPT¹ for unlicensed use as well as frequency range 64-86 GHz for extremely wideband applications [12]. Note that various parts of the 64-86 GHz range are allocated differently in different regions of the world. For instance, 64-71 GHz is set aside for unlicensed use in North America and is under consideration in CEPT (66-71 GHz); 66-76 and 81-86 GHz are under study in ITU-R as possible bands for IMT-2020 to be rat- ified during WRC-19 [12]. In general, the NR UEs do not receive or transmit using the full channel bandwidth of the gNBs, rather they can be assigned to what is referred to as bandwidth parts. While the con- cept does not have any direct RF implications, it is important to note that the gNB and the UE channel bandwidths are defined independently, and the device bandwidth capability does not have to match the gNB channel bandwidth. A unified frame structure is defined in NR that supports TDD, FDD, and half-duplex FDD operations. The duplex scheme is spe- cifically defined for each operating band. Some bands are also designated as SDL or SUL bands to be used in FDD operation. Some of the frequency bands that have been identified for NR deployment are the existing ITU-R IMT2 bands which may have already been accommodating 2G, 3G, and/or 4G [incumbent] deployments. In some regions, a number of The European Conference of Postal and Telecommunications Administrations (CEPT) was established in 1959 and is an organization where policy makers and regulators from 48 countries across Europe collaborate to harmonize telecommunication, radio spectrum, and postal regulations across Europe. The CEPT conducts its work through three autonomous business committees (E

CC, Com-ITU and CERP). The term International Mobile Telecommunications (IMT) is the generic term used by the ITU community to designate broadband mobile systems. It encompasses IMT-2000, IMT-Advanced, and IMT-2020 collectively. New Radio Access RF and Transceiver Design Considerations FR1 (450 MHz to 6 GHz) FR2 (24.25-52.6 GHz) 1.427 1.518 3.4 3.6 4.495 4.8 24.24 31.833.4 50.4 52.6 Frequency (GHz) [IMT-spectrum] [WRC-19 candidate bands] (1.427-1.518 GHz)(3.4-3.6 GHz) (24.25-27.25 GHz) (31.8-33.4 GHz) (37-40.5 GHz) (40.5-42.5 GHz) (42.5-43.5 GHz) (45.5-47 GHz) (47-47.2 GHz) (47.2-50.2 GHz) (50.4-52.6 GHz) (66-76 GHz) (81-86 GHz) 4.2 4.4 Japan 24.25 United States 4.5 4.8 China South Korea Frequency (GHz) 24.25 26.527.528.529.5 Figure 5.1 NR bands and 5G frequency bands under study in ITU-R WRC-19 14,29]. frequency bands are designated and regulated as technology-neutral. which means that the coexistence between different technologies is a requirement for deployment. The capability to operate in this wide range of frequency bands for any cellular system, including NR, has direct implications on the RF requirements. The operators in the same band may deploy NR or other IMT technologies such as LTE [11]. Such coexistence requirements are typically specified within 3GPP (and/or ITU-R), but there may also be regional requirements defined by regulatory bodies that must be satisfied in order to be able to deploy the technology. Mobile operators have diverse spectrum holdings and allocations, which in many cases con- sist of a spectrum block that cannot accommodate one wideband carrier, or the allocation may be non-contiguous, consisting of multiple frequency blocks across multiple bands. In these cases, the NR specifications support carrier aggregation, where multiple carriers within a band, or in multiple bands, can be combined to create effectively wider transmis- sion bandwidths. As shown in Fig. 5.1, the 5G spectrum in various countries is different which makes it difficult to harmonize spectrum and achieve global ro

aming. The LTE and NR coexistence in the same spectrum, which is required for non-standalone and early deployments of 5G systems, makes it possible to deploy NR in the existing frequency allocations. Since the co-channel NR and LTE carriers need to be aligned at sub- carrier level, some restrictions are imposed on the NR channel raster in order to align the position of the NR and LTE carriers. The NR further supports multiple numerologies with 658 Chapter 5 subcarrier spacing (SCS) ranging from 15 to 120 kHz, with direct implications on the time and frequency structures. The subcarrier spacing has certain implications on the RF front- end in terms of the roll-off of the transmitted signal, which impacts the guard bands that are allocated between the transmitted resource blocks and the edge of the frequency band. The NR also supports mixed numerologies on the same carrier, which has further RF implica- tions since the guard bands may need to be different at the two edges of the band. In some cases, there are certain limitations concerning the site where the gNB equipment is deployed. The candidate sites are often shared between operators or an operator deploys multiple technologies at one site, which creates additional requirements for the gNB trans- mitters and receivers to operate in close proximity of other BS transceivers (e.g., blocking effects). The coexistence between operators of TDD systems in the same band is in general provided by inter-operator synchronization in order to avoid interference between downlink and uplink transmissions of different operators. This means that all operators need to have the same uplink/downlink configurations and frame synchronization, which is not by itself an RF requirement, but it is implicitly assumed in the 3GPP specifications. The RF require- ments for unsynchronized systems are inevitably much stricter. The frequency bands are regionally defined, and new bands are added continuously for each generation of mobile systems, which means each new release of 3GPP specificat

ions will incorporate additional bands. Using a release-independent principle, it is possible to design devices based on an early release of 3GPP specifications that support a frequency band added in a later release. The first set of NR bands is defined in 3GPP Rel-15, and additional bands will be added in a release-independent manner. The NR can operate over a wide range of frequencies which include sub-6 GHz and above 6 GHz spectrum. The support of this extremely wide range of frequencies implies that the radio characteristics and operating parameters can significantly vary depending on the fre- quency band and the operating bandwidth; thus it is desirable to be able to configure the radio-related parameters such as subcarrier spacing, OFDM symbol length, cyclic prefix length, channel bandwidth according to the allocated frequency band. The NR specifies two frequency ranges, namely, FR1 and FR2 where each support certain OFDM numerologies. In FR1, which covers 450 MHz to 6 GHz, the subcarrier spacings that can be used for data transmission are 15, 30, and 60 kHz. The FR2 supports subcarrier spacings of 60 and 120 kHz. The UEs are required to support all subcarrier spacings except 60 kHz in FR1. The ratio of cyclic prefix length over OFDM symbol length is the same for all supported subcarrier spacings. Therefore, as the subcarrier spacing increases, the cyclic prefix length decreases, resulting in more susceptibility of NR signals to multipath delay distortion and coverage reduction. The use of larger subcarrier spacings further increases the NR signal tolerance to Doppler shift. It must be noted that the effect of oscillator phase noise is more pronounced in FR2 bands, resulting in inter-carrier interference. The use of larger subcarrier spacing helps mitigate the effects of phase noise. The subcarrier spacing is configured by New Radio Access RF and Transceiver Design Considerations 659 the network using higher layer signaling. The most appropriate subcarrier spacing can be configured according to the deployme

nt scenario. The FR2 signal propagation is character- ized with less diffraction, higher penetration loss, and in general higher path loss. This can be compensated by incorporating more antenna elements at the transmit and receive sides, narrower beams with higher antenna gains, leading to massive MIMO systems. LTE initially supported channel bandwidths up to 20 MHz, and later through the use of car- rier aggregation, it was able to operate in up to 100 MHz bandwidths. The number of com- ponent carriers was later increased to 32 in 3GPP Rel-13, resulting in the maximum operating bandwidth of up to 640 MHz. The Rel-15 NR supports inter-band and intra-band contiguous and non-contiguous carrier aggregation as well as dual connectivity which allows simultaneous communication to NR and LTE base stations. Due to much wider chan- nel bandwidths targeted in NR, the maximum channel bandwidth per component carrier has been increased to 100 MHz in FR1 (using SCSs of 30 and 60 kHz) and 400 MHz in FR2 (using 120 kHz SCS). Support of 400 MHz bandwidth in FR2 is optional; however, the NR devices are required to support up to 200 MHz in FR2. The NR new bands and the cor- responding channel bandwidths and SCSs are shown in Table 5.1 [29]. The LTE spectrum utilization (defined as the ratio of the transmission bandwidth over the channel bandwidth) was 90% and guard bands were provisioned on both sides of the trans- mission bandwidth to protect communication in the adjacent channels by limiting the adja- cent channel interference. In NR, the spectrum utilization has been increased to 98% (depending on transmission bandwidth), as a result of time and frequency-domain prepro- cessing of the OFDM signal (i.e., windowing and spectral shaping). While the size of the guard bands have been reduced in NR due to increased spectrum utilization, the out-of-band (OOB) emission and the permissible leakage power requirements have remained the same as LTE. Table 5.1: NR new bands and channel bandwidths ("x" configurable and optional) [5,29]. Frequ

ency Frequency Band Subcarrier Component Carrier Bandwidth (MHz) Range Frequency Spacing Number (kHz) (GHz) n257/ 24-40 n260/ 660 Chapter 5 BW channel- channel bandwidth (MHz) Guard band Transmission bandwidth configuration NRB [RB] Guard band Transmission bandwidth [RB] Spectrum Resource block Active resource blocks Note: guard bands can be asymmetric. emission mask Figure 5.2 Illustration of channel bandwidth, transmission bandwidth, and guard bands [5]. The channel bandwidth of a gNB (BW channel) supports transmission on a single (NR) uplink or downlink RF carrier. Different UE channel bandwidths may be supported within the same gNB channel bandwidth for bidirectional communication with the UE. The relative location of the UE channel bandwidth is flexible and within the gNB channel bandwidth. The gNB is able to transmit/receive to/from one or more UE bandwidth parts that are smal- ler than or equal to the total number of carrier resource blocks on the RF carrier. The rela- tionship between the channel bandwidth, guard band, and the transmission bandwidth is shown in Fig. 5.2. The transmission bandwidth configuration in terms of the number of physical resource blocks (NRB) for each BS channel bandwidth and SCS was given in Table 3.9, for FR1 and FR2. The minimum guard band for each BS channel bandwidth and SCS is shown in Table 5.2 for FR1 and FR2. The minimum guard band shown for SCS of 240 kHz is only applicable when an SS/PBCH block (SSB) with SCS of 240 kHz is placed adjacent to the edge of the BS channel bandwidth within which the SSB is located; other- wise, 240 kHz subcarrier spacing is not used for any other configurations. The number of RBs configured within any BS channel bandwidth will guarantee that the minimum guard band shown in Table 5.2 is satisfied. If multiple numerologies are multiplexed over the same OFDM symbol, the minimum guard band on each side of the carrier is the guard band applied at the configured BS channel bandwidth for the numerology that is transmitted/received adjacent to the g

uard band. In FR1, if multiple numerologies are multiplexed over the same OFDM symbol and the BS channel bandwidth is greater than 50 MHz, then the guard band that is inserted adjacent to 15 kHz SCS is the same as the guard band defined for 30 kHz SCS for the same BS channel bandwidth. In FR2, if multiple numerologies are multiplexed over the same OFDM symbol Table 5.2: Minimum guard band (kHz) for FR1 and FR2 [5]. Subcarrier Bandwidth (MHz) Spacing (kHz) 242.5 312.5 382.5 452.5 522.5 592.5 552.5 692.5 215,560 662 Chapter 5 and the BS channel bandwidth is larger than 200 MHz, then the guard band inserted adja- cent to 60 kHz SCS is the same as the guard band defined for 120 kHz SCS for the same BS channel bandwidth [5]. For each BS channel bandwidth and each numerology, the BS transmission bandwidth configuration is required to satisfy the minimum guard band requirement. The common resource blocks (see Section 3.7.1) are specified for each numerology and the starting point of the associated transmission bandwidth configuration on the common resource block grid for a given channel bandwidth is indicated by an offset to reference point A in the unit of the numerology. For each numerology, all UE transmission bandwidth configurations that are indicated to the UEs by the serving gNB through higher layer parameter carrierBandwidth must be located within the BS transmission bandwidth configuration [5]. In carrier aggregation scenarios, the transmission bandwidth configuration is defined per component carrier. An aggregated BS channel bandwidth and guard bands are defined for intra-band contiguous carrier aggregation. The aggregated BS channel bandwidth (BWchannel_CA) is defined as BW channel_CA = fedge-high -fedge-low (MHz). As shown Fig. 5.3, the lower bandwidth edge fedge-low and the upper bandwidth edge fedge-high of the aggregated BS channel bandwidth are used as frequency reference points for transmitter and receiver requirements, which are defined as fedge-low = fc-low - foffset-low Aggregated BS channel bandwidt

h BWChannel-CA Lowest carrier bandwidth Highest carrier bandwidth configuration (RB) configuration (RB) Lower Upper foffset-low foffset-high Spectrum edge-low C-high edge-high Sub-block gap Sub-block Sub-block RF bandwidth Operating band Figure 5.3 Illustration of aggregated BS channel bandwidth for intraband carrier aggregation [5]. New Radio Access RF and Transceiver Design Considerations 663 fedge-high = fc-high +foffset-high. The lower and upper frequency offsets depend on the trans- mission bandwidth configurations of the lowest and highest assigned edge component carrier and are defined as foffset-low = (12NRB-low + 1)SCSlow/2 + WGB (MHz) and foffset-high = (12NRB-high - 1)SCShigh/2 + WGB(MHz), where WGB denotes the lower or upper minimum guard band defined for the lowest and highest assigned component car- rier. In the latter equations, the parameters NRB-low and NRB-high are the transmission bandwidth configurations for the lowest and highest assigned component carrier, SCSlow and SCShigh denote the subcarrier spacing for the lowest and highest assigned component carrier, respectively [5]. It can be shown that the minimum guard band can be defined as WGB=BWchannel/2-SCS(12NRB = + 1)/2, where NRB is the maximum number of resource blocks that can fit into the BS channel bandwidth and SCS denotes the subcarrier spac- ing. Note that an extra SCS/2 guard band is inserted on each side of the carrier due to the relation to the RF channel raster, which has a subcarrier-level granularity and is defined independent of the actual spectrum blocks. Therefore, it may not be possible to place an RF carrier exactly in the center of a spectrum block and an extra guard band would be required to ensure that the RF requirements can be met [10]. Some spectrum allocations may consist of fragmented blocks of spectrum. In the intra-band non-contiguous case, the BS transmits and receives over an RF bandwidth that is split in two (or more) separate subblocks with a subblock gap in between (see Fig. 5.3). A subblock is defined as o

ne contiguous allocated block of spectrum for transmission and reception by the same BS. The lower subblock edge of the subblock bandwidth (BW channel-block) defined as fedge-block-low = fc-block-low foffset-low. The upper subblock edge of the subblock bandwidth is defined as fedge-block-high fc-block-high -foffset-high- The subblock bandwidth channel-block is defined as BW channel-block = fedge-block-high -fedge-block-low (MHz). The lower and upper frequency offsets foffset-block-low and foffset-block-high depend on the transmission bandwidth configurations of the lowest and highest assigned edge component carriers within a subblock which are defined as foffset-block-low = (12NRB-low 1)SCSlow/2 + WGB-low (MHz) and foffset-block-high = (12NRB-high 1)SCShigh/2 + WGB-high (MHz), where NRB-low and NRB-high are the transmission bandwidth configurations for the lowest and highest assigned compo- nent carrier within a subblock, respectively. In the latter equations, SCSlow and SCShigh denote the subcarrier spacing for the lowest and highest assigned component carrier within a subblock, respectively; WGB-low and WGB-high are the minimum guard bands for the lowest and highest assigned component carriers, respectively. The subblock gap size between two consecutive subblocks Wgap is defined as Wgap = fedge-block-low(n+1) -fedge-blockn-high(n (MHz). Because the subblock gap starts from the inner edge of the channel bandwidth and not the center of the channel bandwidth, the subblock gap width is independent of the component carriers' channel bandwidth [5]. The frequency spacing between RF carriers depends on the deployment scenario, the size of the frequency block available and the BS channel bandwidths. The nominal channel 664 Chapter 5 spacing DF between two adjacent NR carriers is defined based on the frequency range and channel raster. For NR FR1 operating bands with 100 kHz channel raster, DF = (BW channel(2))/2. For NR FR1 operating bands with 15 kHz channel = FR2 operating bands with 60 kHz channel raster, DF = N chan

nel(2) / 2 - 2 - - 0, 20} kHz where BW channel(1) and BW channel(2) are the BS channel bandwidths of the two NR RF carriers. The channel spacing can be adjusted depending on the channel raster in order to optimize the system performance in a particular deployment scenario [5]. The channel spacing between adjacent component carriers for intra-band contiguously aggregated carriers is a multiple of least common multiple of channel raster and subcarrier spacing. The nominal channel spacing DF between two adjacent aggregated NR carriers is defined as follows [5]: For NR operating bands with 100 kHz channel raster (MHz) For NR operating bands with 15 kHz channel raster 0.015 (MHz) and max(m1,m2) For NR operating bands with 60 kHz channel raster (MHz) and n = max (111, In the above expressions, BW channel(1) and BW channel(2) represent the BS channel band- widths of the two NR component carriers with values in MHz; WGBchannel() is the minimum guard band of the ith channel; 11 and H2 denote the subcarrier spacing configurations of the component carriers. The channel spacing for intra-band contiguous carrier aggregation can be adjusted to any multiple of least common multiple of the channel raster and SCS less than the nominal channel spacing to optimize performance in a particular deployment scenario. For intra-band non-contiguous carrier aggregation, the channel spacing between two NR component carriers in different subblocks must be larger than the nominal chan- nel spacing [5]. New Radio Access RF and Transceiver Design Considerations 5.2 Base Station Transceiver RF Characteristics and Requirements 5.2.1 General Base Station RF Requirements The RF requirements in general are defined at the BS antenna connector. Those requirements are referred to as conducted requirements and are typically specified as absolute or relative power levels measured at the antenna connector. The OOB emission limits are often defined as conducted requirements. In active antenna systems, the RF requirements are defined as radiated requirement

, which are measured over the air in the far-field of the antennas; thus they include the antenna patterns and directivity effects [1]. In OTA test procedures, the spatial characteris- tics of the BS including the antenna system are evaluated. In base stations equipped with AAS, where the active parts of the transceiver are integrated with the antenna system, it is not always possible to conduct the measurements at the antenna connector. For this reason, 3GPP Rel-13 specified the RF requirements for the AAS base stations in a set of separate RF specifications that are applicable to LTE and the previous generations. The radiated RF requirements and OTA testing are being specified for the new radio for the FR1 and FR2 bands, which have bor- rowed a large portion of the previously developed AAS specifications. Note that the term AAS is not used within the NR base station RF specifications, rather the requirements are specified for different BS types. The AAS-type BS requirements are based on the generalized AAS BS radio architecture shown in Fig. 5.4 [5]. The architecture consists of a transceiver unit array that is connected to a composite antenna structure that contains a radio distribution network and an antenna array. The transceiver unit array comprises a number of transmitter and receiver units, which are connected to the composite antenna structure via a number of connectors on the transceiver array boundary (TAB). The TAB connectors correspond to the antenna connectors on a non- AAS BS and serve as a reference point for the conducted requirements. The radio distribution network is a passive unit which distributes or aggregates the transmitter outputs or the receiver inputs to the corresponding antenna elements, respectively. It must be noted that the actual implementation of an AAS BS may be different in terms of the physical location of different parts, array geometry, type of antenna elements, etc. Based on the architecture shown in Fig. 5.4, two types of requirements can be defined. Conducted requirements

are defined for each RF characteristic at an individual or a group of TAB connectors. The con- ducted requirements are defined such that they are equivalent to the corresponding conducted requirement of a non-AAS BS, which implies that the performance of the system or the impact on other systems is expected to be the same. Radiated requirements, on the other hand, are defined based on OTA measurements conducted in the far-field of the antenna sys- tem. Since the spatial direction becomes relevant in this case, it is detailed for each require- ment how it applies. The radiated requirements are defined with reference to a radiated interface boundary (RIB) in the far-field region of the antenna array [5,10]. 666 Chapter 5 Transceiver array boundary Radiated interface boundary (RIB) Transceiver unit array (TRXUA) 1 to M Transceiver unit 1 Radio Antenna Transceiver unit 2 distribution network array Transceiver unit M Composite antenna Transceiver array boundary connector (TAB) Figure 5.4 Radiated and conducted reference points for BS type 1-H [5]. BS, Base station. Rayleigh region Fresnel region Fraunhofer region (reactive near field) (radiative near field) (far field) Antenna/ Antenna array d < 0.62 D3/2 d < 2D 12 d < 2D 2 Figure 5.5 Illustration of the radiation regions of an antenna element/array. In order to determine an appropriate point in the antenna array far-field region for conducting the OTA measurements, one must understand the radiation characteristics of the antenna sys- tems and be able to define the minimum distance dmin in the far-field radiation region of an antenna array. In the theory of electromagnetics and antennas, the radiation regions surround- ing an antenna element/array can be divided into three regions as follows (see Fig. 5.5): Reactive near-field region: In the immediate vicinity of the antenna, we have the reac- tive near-field. In this region, the fields are predominately reactive fields, which means the electric (E) and the magnetic (H) fields are out of phase by 90 degrees relative

to each other. Note that for propagating or radiating fields, the fields are orthogonal and are in-phase. The boundary of this region is typically given as d < 0.62 D3 1, where New Radio Access RF and Transceiver Design Considerations 667 D is the maximum linear dimension of an antenna element/array, 1 denotes the wave- length, and d is the distance from the antenna(s). Radiating near-field (Fresnel) region: The radiating near-field or Fresnel region is the region between the near- and far-fields. In this region, the reactive fields are not domi- nant, and the radiating fields begin to emerge. However, unlike the far-field region, the shape of the radiation pattern may vary noticeably with distance. This region is often identified by 0.62(/D3/X<d<2D2/A. Note that depending on the values of d and the wavelength, this field may or may not exist. Far-field (Fraunhofer) region: The far-field is a region relatively far from the antenna element/array, where the shape of radiation pattern does not vary with distance, although the fields are attenuated proportional to d-1, the power density degrades by d-2. This region is dominated by radiated fields, with the E- and H-fields in a plane orthogonal to each other and to the direction of propagation which is the main charac- teristic of plane waves. If the maximum linear dimension of an antenna is D, the follow- ing three conditions must all be satisfied SO that a point can be in the far-field region: d > 2D2/1, d D and d 1. The first two inequalities ensure that the power radiated in a given direction from distinct parts of the antenna is approximately parallel. This helps ensure that the fields in the far-field region can be characterized as plane waves. Near a radiating antenna, there are reactive fields that typically have the E-fields and H-fields diminish with distance as d-2 or d-3. The third inequality ensures that these near-fields are disappeared, and we are left with the radiating fields, which diminish with distance as d-1. The NR base stations can be categorize

d into three main configurations from RF require- ments perspective as follows [5]: BS type 1-C: An NR BS operating in FR1 which is connected to the antennas via coax- ial cables. The RF requirements for this BS type only consist of conducted requirements defined at individual antenna connectors. For BS type 1-C, the requirements are applied at the BS antenna connector (port A). If any external apparatus such as an amplifier, a filter, or the combination of such devices is used, the RF requirements apply at the far end antenna connector (port B). BS type 1-H: An NR BS operating in FR1 consisting of an integrated AAS with RF trans- ceiver connected to antennas using a TAB connector. The requirement set for this BS type consists of conducted requirements defined at individual TAB connectors and OTA requirements defined at the RIB. For BS Type 1-H, the requirements are defined for two points of reference, signified by radiated requirements and conducted requirements. BS type 1-0: A connector-less AAS-type NR BS operating in FR1 whose RF require- ment set only consist of OTA requirements defined at the RIB. BS type 2-O: A connector-less AAS-type NR BS operating in FR2 whose RF require- ment set only consists of OTA requirements defined at the RIB. 668 Chapter 5 For BS Type 1-O and BS Type 2-O, the radiated characteristics are defined over the air, where the operating band-specific radiated interface is referred to as the RIB. The radiated requirements are also referred to as OTA requirements. The (spatial) characteristics in which the OTA requirements apply are detailed for each requirement. Fig. 5.6 illustrates the NR BS configurations and the reference points for conducting connected or OTA measure- ments. Note that one of the RF configurations for FR1 does not require connectors between the RF transceivers and antennas according to BS Type 1-O; thus smaller equipment size and improved power efficiency can be expected relative to the LTE transceivers with AAS specified in 3GPP Rel-13. While the operation in FR2 has

the advantage of wideband trans- mission in high-frequency bands in terms of RF configuration, the higher frequencies result in larger power losses at the connectors and cables; increased path loss and reduced cover- age due to lower power density over wider channels. Therefore, higher antenna gains are necessary to compensate for the losses and to maintain a certain link budget and coverage. Since it would be more difficult to design and implement RF signal transceivers and anten- nas with high density in FR2, if conventional RF configuration with connectors is used, BS type 2-O RF configuration without connectors is defined for FR2 operation. The BS Type 2-O would allow implementation of beamforming over wide channel bandwidths to main- tain coverage and to achieve high spectral efficiencies [29]. The RF performance requirements for BS types 1-C and 1-H are based on LTE-advanced specifications [11], when NR radio parameters are applied. However, BS types 1-O and 2-O have integrated radio transceivers and antennas with no connectors to conduct measure- ments; thus OTA specifications have been extended such that in the overall RF performance specifications, a reference point in the radiated space, referred to RIB, can be defined. The output power for various BS classes and types is shown in Table 5.3. It must be noted that the factor 10 log(NTxU-counted) is used to derive the rated carrier output power from output power per TAB connector for BS Type 1-H and no upper limits for output power has been specified for BS Type 2-O in 3GPP Rel-15. In addition to the equivalent isotropic radiated power (EIRP) and the equivalent isotropic sensitivity (EIS) including the antenna characteristics in the beam direction, which were specified in 3GPP Rel-13 LTE, the total radiated power (TRP) is introduced as a new metric in NR RF specifications (Fig. 5.7). The TRP definition makes it possible to specify OTA requirements for power-related RF performance requirements such as the BS output power and spurious emissions. Fig. 5.7 i

llustrates the visualization of the EIRP, EIS, and TRP definitions. The main BS RF performance specifications of the LTE and NR in FR1 and FR2 have been summarized and compared in Table 5.4. The RF specifications in FR1 are based on LTE specifications with the maximum channel bandwidth of 100 MHz. In FR2, the NR radio specifications support wider bands, lower latency, and faster response with maximum Base station type BS type 1-C BS type 1-H BS type 1-0/2-O (equivalent to legacy BS) (equivalent to AAS BS in LTE Rel-13) (new BS types for NR) Frequency range FR1/FR2 Antenna Transceiver unit 1 Transceiver unit 1 Transceiver unit 1 Transceiver unit 2 Transceiver unit 2 Transceiver unit 2 Base station Antenna connector configuration (reference point Antenna for conducted Antenna requirements) Transceiver unit M connector Transceiver unit M Transceiver unit M Coaxial cable TAB connector (Reference point for conducted requirements and partial OTA requirements) Figure 5.6 NR base station configurations 5,29]. 670 Chapter 5 Table 5.3: Output power per base station class and type (dBm) [5,14]. BS Type 1-C BS Type 1-H BS Type 1-O The Sum of Rated Carrier The Rated Carrier Output Power for All TAB The Rated Carrier Rated Carrier TRP Output Power per Connectors for a Single Output Power per Output Power Antenna Connector Carrier TAB Connector Declared per RIB No upper limit for wide area base station area BS Medium < 38 + 10 log(NTxU-counted) range BS Local < 24 + 10 log(NTxU-counted) area BS Equipment under test Equipment under test Emitted beam Emitted beam EIRP/EIS (regulated by beam direction, including antenna performance) (Regulated using entire sphere surface) EIRP/EIS definition TRP definition Figure 5.7 Illustration of the RF performance requirements for the NR base station and mobile station 29]. Table 5.4: Comparison of main base station RF performance specifications [5,29]. NR FR1 NR FR2 BS Type 1-O BS Type 2-O Maximum channel bandwidth (MHz) Transmitter transient period (us) ACLR (dB) NF (dB) Transmit power devia