Datasets:

eswardivi

/

books

like 0

state machine and state transitions between NR/5GC, LTE/EPC, and LTE/5GC [1]. handled by the device through cell reselection, while in the connected state, the mobility is managed by the network based on measurements. Fig. 2.24 illustrates an overview of UE RRC state machine and state transitions in NR. A UE has only one RRC state in NR at a given time. The UE is either in RRC_CONNECTED state or in RRC_INACTIVE state when an RRC connection has been established; otherwise, if no RRC connection is established, the UE is in RRC_IDLE state. The RRC states can fur- ther be characterized as follows (see the summary in Table 2.2): RRC_IDLE: In this state, a UE-specific DRX may be configured by upper layers. A UE-controlled mobility based on network configuration will be used. The UE monitors short messages transmitted with P-RNTI over downlink control channel. It also moni- tors the PCH for CN paging using 5G-S-TMSI and performs neighboring cell measure- ments and cell (re) selection. It further acquires SI and can send SI request (if configured). RRC_INACTIVE: In this state, a UE-specific DRX may be configured by upper layers or by RRC layer. A UE-controlled mobility based on network configuration is used. The UE stores the inactive AS context and an RNA is configured by the RRC sublayer. The UE monitors short messages transmitted with P-RNTI over downlink control chan- nel. It monitors the PCH for CN paging using 5G-S-TMSI and RAN paging using full I-RNTI and performs neighboring cell measurements and cell (re) )selection. The UE fur- ther periodically performs RNA updates (RNAUs) when moving outside the configured RNA. It also acquires SI and can send SI request (if configured). RRC_CONNECTED: In this state, the UE stores the AS context. The network transfers unicast data to/from the UE. At the lower layers, the UE may be configured with a UE-specific DRX. For carrier-aggregation-capable UEs, the network may use one or 246 Chapter 2 more SCells, aggregated with the SpCell to increase operation bandwidth. For the UEs

supporting dual connectivity, the network may use one SCG, aggregated with the MCG, for increased operational bandwidth. Network-controlled mobility within NR and to/ from E-UTRA is used in this mode. The UE monitors short messages transmitted with P-RNTI over downlink control channel. It further monitors control channels associated with the shared data channel to determine if data is scheduled for it. The UE provides channel quality and feedback information to the gNB and conducts neighbor cell mea- surements and measurement reporting and acquires the SI. 2.3.3.1 Idle Mode Procedures The RRC_IDLE state and RRC_INACTIVE state procedures can be divided into three pro- cesses, namely, PLMN selection, cell selection/reselection, and location registration and RNA update. The PLMN selection, cell reselection procedures, and location registration are common for both RRC_IDLE state and RRC_INACTIVE state, whereas RNA update is only applicable to RRC_INACTIVE state. When the UE selects a new PLMN, it transitions from RRC_INACTIVE to RRC_IDLE state. When a UE is powered on, a PLMN is selected by NAS and a number of RATs associated with the selected PLMN are identified for cell selection. The NAS provides a list of equivalent PLMNs that the AS must use for cell selec- tion/reselection. During cell selection the UE searches for a suitable cell within the selected PLMN. The UE would select a cell, if certain criteria are met. This procedure is known as camping on the cell in 3GPP terminology. The UE then registers with the network by means of NAS registration procedure in the tracking area of the selected cell. A successful location registration would make the selected PLMN as the registered PLMN. If the UE finds a more suitable cell, based on the cell reselection criteria, it reselects that cell and camps on it. If the new cell does not belong to at least one tracking area to which the UE is registered, another location registration is performed. In RRC_INACTIVE state, if the new cell does not belong to the configured RNA,

an RNA update procedure is performed. The UE usually searches for higher priority PLMNs at regular time intervals and continues to search for a more suitable cell, if another PLMN has been selected by NAS. If the UE loses the coverage of the registered PLMN, either a new PLMN is automatically selected (automatic mode), or the user is notified of the available PLMNs SO that a manual selection can be performed (manual mode). The purpose of camping on a cell in RRC_IDLE or RRC_INACTIVE state is to enable the UE to receive SI from the PLMN, when registered. If the network needs to send a message or deliver data to the registered UE, it knows the set of tracking areas (in RRC_IDLE state) or RNA (in RRC_INACTIVE state) in which the UE is camped. The network can then send a paging message to the UE on the control channels of all cells in the corresponding set of (tracking) areas. It further enables the UE to receive ETWS and CMAS notifications. During the PLMN selection, the UE scans all RF channels in the NR bands according to its New Radio Access Layer 2/3 Aspects and System Operation 247 capabilities to find available PLMNs. On each carrier, the UE searches for the strongest cell and acquires the SI in order to find which PLMN(s) the cell belongs to. If the UE can detect one or several PLMN identities in the strongest cell, each detected PLMN is reported to the NAS as a high-quality PLMN provided that the measured reference signal received power value is greater than or equal to 110 dBm [12]. As we mentioned earlier, the PLMN selection in NR is based on the 3GPP PLMN selection, and that is, cell selection is required upon transition from RM-DEREGISTERED to RM-REGISTERED, from CM-IDLE to CM-CONNECTED, and from CM-CONNECTED to CM-IDLE based on the following principles [8]: The UE NAS layer identifies a PLMN and equivalent PLMNs. Cell selection is always based on cell-defining SSBs (CD-SSBs)9 located on the chronization raster. The UE searches the designated NR frequency bands and for each carrier frequency identifies

the strongest cell consistent with the CD-SSB. It then detects the SI broadcast from that cell to identify its PLMN(s). The UE may conduct a full search in initial cell selection or make use of the stored information to shorten the search, that is, stored information cell selection. The UE searches for a suitable cell, and if it is not able to identify a suitable cell, it may proceed with an acceptable cell. When a suitable cell or an acceptable cell is found, the UE camps on that cell and starts the cell reselection procedure. A suitable cell is one for which the measured cell attributes satisfy the cell selection criteria; the cell PLMN is the selected PLMN, registered, or an equivalent PLMN; the cell is not barred or reserved, and the cell is not part of a tracking area which in the list of forbidden tracking areas for roaming. An acceptable cell is one for which the measured cell attributes satisfy the cell selection criteria and the cell is not barred. Upon transition from RRC_CONNECTED or RRC_INACTIVE to RRC_IDLE, the UE should camp on a cell following cell selection depending on the frequency assigned by the RRC sublayer in the state transition message. The UE should attempt to find a suitable cell in the manner described for stored infor- mation or initial cell selection. If no suitable cell is found on any frequency or RAT, the UE should attempt to find an acceptable cell. In multi-beam operation, the cell quality is derived amongst the beams corresponding to the same cell. Within the frequency span of a carrier, multiple SSBs can be transmitted. The PCIs of SSBs transmitted in different frequency locations do not have to be unique, that is, different SSBs in the frequency domain can have different PCIs. However, when an SSB is associated with an RMSI, the SSB corresponds to an individ- ual cell, which has a unique NR cell global identifier (NCGI). Such an SSB is referred to as CD-SSB. A PCell is always associated with a CD-SSB located on the synchronization raster. 248 Chapter 2 A UE in RRC_IDLE may per

form cell reselection according to the following procedure [8]: Cell reselection is always based on CD-SSBs located on the synchronization raster. The UE measures the attributes of the serving and neighbor cells to facilitate the rese- lection process. The UE would only need information on the carrier frequencies of the inter-frequency neighbor cells, when conducting search and measurement. Cell reselection identifies the cell that the UE should camp on. It is based on cell rese- lection criteria which involves measurements conducted on the serving and the neighbor cells. It must be noted that intra-frequency cell reselection is based on ranking of the cells and inter-frequency cell reselection is based on absolute priorities, where a UE would camp on the highest priority frequency available. A neighbor cell list (NCL) can be provided by the serving cell to facilitate cell selection in specific cases for intra- and inter-frequency neighbor cells. The NCL contains cell- specific cell reselection parameters (e.g., cell-specific offset) for specific neighbor cells. Black lists can be provided to prevent the UE from reselecting to specific intra- and inter-frequency neighboring cells. In multi-beam operations, the cell quality is derived amongst the beams corresponding to the same cell. 2.3.3.2 Inactive Mode Procedures RRC_INACTIVE is a state where a UE remains in CM-CONNECTED state while roaming within an area configured by NG-RAN known as RNA without notifying NG-RAN. In RRC_INACTIVE state, the last serving gNB maintains the UE context and the UE-associated NG connection with the serving AMF and user-plane function (UPF). If the last serving gNB receives downlink data from the UPF or downlink UE-associated signaling from the AMF (except the UE Context Release Command message) while the UE is in RRC_INACTIVE state, it pages the UE in the cells corresponding to the RNA and may send an XnAP RAN Paging 10 to neighbor gNB(s), if the RNA includes cells of neighboring gNB(s). Upon receiving the UE Context Release Command

message while the UE is in RRC_INACTIVE state, the last serving gNB may send the paging message in the cells cor- responding to the RNA and may send an XnAP RAN Paging to neighbor gNB(s), if the RNA includes the cells of neighbor gNB(s). Upon receiving the NG RESET message while the UE is in RRC_INACTIVE state, the last serving gNB may page the involved UE(s) in The purpose of the RAN Paging procedure is to enable the NG-RAN node to request paging of a UE in the NG-RAN node2. The procedure uses non-UE-associated signaling. The RAN paging procedure is triggered by the NG-RAN node by sending the RAN paging message to the NG-RAN node2, in which the necessary information such as UE RAN Paging identity is provided. New Radio Access Layer 2/3 Aspects and System Operation 249 the cells corresponding to the RNA and may send an XnAP RAN Paging to neighbor gNB (s), if the RNA includes the cells of neighbor gNB(s) [8]. The AMF provides to the NG-RAN node the CN assistance information to assist the NG-RAN node's decision on whether the UE can be moved to the RRC_INACTIVE state. The CN assistance information includes the registration area configured for the UE, the periodic registration update timer, and the UE identity index value, and may further include the UE-specific DRX, an indication if the UE is configured with mobile initiated connection only mode by the AMF, and the expected UE behavior. The UE registration area is consid- ered by the NG-RAN node, when configuring the RNA. The UE-specific DRX and UE identity index value are used by the NG-RAN node for RAN paging. The periodic registra- tion update timer is taken into account by the NG-RAN node to configure periodic RNA update timer. The NG-RAN node further considers the expected UE behavior to assist the UE RRC state transition decision [8]. During transition to RRC_INACTIVE state, the NG-RAN node may configure the UE with a periodic RNA update timer value. If the UE attempts to access a gNB other than the last serving gNB, the receiving gNB triggers the XnAP retrie

receiving gNB triggers the XnAP retrieve UE context procedure to obtain the UE con- text from the last serving gNB and may decide to move the UE back to RRC_INACTIVE state, RRC_CONNECTED state, or RRC_IDLE state. In case of peri- odic RNA update, if the last serving gNB decides not to relocate the UE context, it fails the retrieve UE context procedure and directly moves the UE back to RRC_INACTIVE state or to RRC_IDLE state by an encapsulated RRCRelease message. Table 2.3 provides the functional split between the UE NAS and AS procedures in RRC_IDLE and RRC_INACTIVE states. The UE may use DRX in RRC_IDLE and RRC_INACTIVE states in order to reduce power consumption. The UE monitors one paging occasion (PO) per DRX cycle. A PO is a set of Table 2.3: Functional split between access stratum and non-access stratum in RRC IDLE and RRC_INACTIVE states [12]. RRC_IDLE and UE Non-access Stratum UE Access Stratum RRC_INACTIVE State Procedure PLMN selection Maintain a prioritized list of PLMNs Search for available PLMNs, if the associated RAT(s) are Select a PLMN using automatic or manual mode set for the PLMN, search among those RAT(s) Request AS to select a cell belonging to this PLMN. For Perform measurements to support PLMN selection each PLMN, associated RAT(s) may be set Synchronize to a broadcast channel to identify PLMNs Evaluate reports of available PLMNs from AS for PLMN Report available PLMNs with the associated RAT(s) to selection NAS on request from NAS or autonomously Maintain a list of equivalent PLMN identities Cell selection Control cell selection by indicating RAT(s) associated Perform the required measurements to support cell with the selected PLMN to be used initially in the search selection of a cell in the cell selection process Detect and synchronize to a broadcast channel Maintain a list of forbidden tracking areas and provide the Receive and process broadcast information list to AS Forward NAS system information to NAS Search for a suitable cell. The cells broadcast one or more PLMN identity in the s

ystem information. Search among the associated RATs for that PLMN Respond to NAS whether such cell is found If a cell is found which satisfies cell selection criteria, camp on that cell Cell reselection Maintain a list of equivalent PLMN identities and Perform the required measurements to support cell provide the list to AS reselection Maintain a list of forbidden tracking areas and provide the Detect and synchronize to a broadcast channel list to AS Receive and process broadcast information Forward NAS system information to NAS Change cell if a more suitable cell is found Location Register the UE as active after power up Report registration area information to NAS registration Register the UE's presence in a registration area regularly or when entering a new tracking area De-register UE when shutting down Maintain a list of forbidden tracking areas RAN notification Register the UE's presence in a RAN-based notification area update area and periodically or when entering a new RNA New Radio Access Layer 2/3 Aspects and System Operation 251 PDCCH monitoring occasions and can consist of multiple time slots (e.g., subframe or OFDM symbols) where paging DCI can be sent. One paging frame (PF) is one radio frame and may contain one or multiple PO(s) or starting point of a PO. In multi-beam operations, the UE can assume that the same paging message is repeated in all transmitted beams, and thus the selection of the beam(s) for the reception of the paging message is up to UE imple- mentation. The paging message is the same for both RAN-initiated paging and CN-initiated paging. The UE initiates RRC Connection Resume procedure upon receiving RAN-initiated paging. If the UE receives a CN-initiated paging in RRC_INACTIVE state, the UE transi- tions to RRC_IDLE state and informs the NAS. The PF and PO for paging are determined by the following expressions. The SFN for the PF is determined by (SFN PF_offset)modT = (T/N)(UE_IDmodN) where index is indicating the index of the PO is determined by is = UE_ID/N mod The parameters of

the latter equations are defined as follows: T denotes the DRX cycle of the UE where T is determined by the short- est of the UE-specific DRX value, if configured by RRC or upper layers and a default DRX value broadcast in SI. If UE-specific DRX is not configured by RRC or by upper layers, the default value is applied; N is the number of total PFs in T; Ns denotes the number of POs for a PF; PF_offset is the offset used for PF determination; and UE_ID = 5G-S-TMSImod1024 [12]. A UE in RRC_INACTIVE state performs cell reselection similar to the procedure earlier defined for the RRC_IDLE state. The UE in the RRC_INACTIVE state can be configured by the last serving NG-RAN node with an RNA, where the RNA can cover a single or mul- tiple cells and is contained within the CN registration area, as well as an RNA update that is periodically sent by the UE and is also sent when the cell reselection procedure of the UE selects a cell that does not belong to the configured RNA [8]. 2.3.3.3 Connected Mode Procedures In the RRC_CONNECTED state, the device has a connection established to the network. The goal of connected-state mobility is to ensure that this connectivity is sustained without interruption or noticeable degradation as the device moves across the network. To satisfy this goal, the device continuously searches for and conducts measurements on new cells both at the current carrier frequency (intra-frequency measurements) and at different car- rier frequencies (inter-frequency measurements) that the device has been configured to do. Such measurements are conducted on the SSB in the same way as for initial access and cell selection/reselection in idle and inactive modes. However, the measurements can also be conducted on configured CSI-RS. In the connected mode, the handover is network- controlled, and the UE does not make any decision on handover to a different cell. Based on different triggering conditions such as the relative power of a measured SSB relative to that of the current cell, the device reports the resu

lt of the measurements to the network. The network then makes a decision as to whether the device has to be handed-over to a 252 Chapter 2 new cell. It should be noted that the reporting is provided through RRC signaling and not layer 1 measurement and reporting framework used for beam management. Apart from some cases in small cell network architectures where the cells are relatively synchronized, the device must perform a new uplink synchronization with respect to the target cell prior to handover. To obtain synchronization to a new cell, the UE has to perform a contention- free random-access procedure using resources specifically assigned to the device with no risk of collision with the goal of attaining synchronization to the target cell. Thus, only first two steps of the random-access procedure are needed which includes the preamble transmission and the corresponding random-access response providing the device with updated transmission timing [17]. In RRC_CONNECTED state, a network-controlled mobility scheme is used which has two variants: cell-level mobility and beam-level mobility. In cell-level mobility explicit RRC signaling is used to trigger a handover. For inter-gNB handover, the signaling procedures consist of four components as follows [8]: 1. The source gNB initiates handover and issues a Handover Request message over Xn interface. 2. The target gNB performs admission control and provides the RRC configuration as part of the Handover ACK message. 3. The source gNB provides the RRC configuration to the UE in the Handover Command. The Handover Command message includes at least the cell ID and all information required to access the target cell SO that the UE can access the target cell without detect- ing its SI. In some cases, the information required for contention-based and contention- free random-access procedure can be included in the Handover Command message. The access information to the target cell may include beam-specific information. 4. The UE moves the RRC connection to the target gNB and r

eplies with a Handover Complete message. The handover mechanism triggered by RRC signaling requires the UE to at least reset the MAC entity and reestablish RLC entity. The NR supports RRC-managed handovers with and without PDCP entity reestablishment. For DRBs using RLC-AM mode, PDCP can either be reestablished along with security key change or initiate a data recovery proce- dure without key change. For DRBs using RLC-UM mode and for SRBs the PDCP entity can be reestablished either together with security key change or to remain as it is without key change. Data forwarding, in-sequence delivery, and duplication avoidance during handover can be guaranteed, when the target gNB uses the same DRB configuration as the source gNB. The NR further supports timer-based handover failure procedure where the RRC connection reestablishment procedure is used for recovering from handover failure. New Radio Access Layer 2/3 Aspects and System Operation 253 In beam-level mobility, explicit RRC signaling is not required to trigger handover. The gNB provides the UE via RRC signaling the measurement configuration containing configura- tions of SSB/CSI-RS resources and resource sets, reports, and trigger states for triggering channel and interference measurements and reports. The beam-level mobility is performed at lower layers by means of physical layer and MAC layer control signaling, and the RRC is not required to know which beam is being used at any given time. The SSB-based beam- level mobility is based on the SSB associated with the initial DL BWP and can only be con- figured for the initial DL BWPs and for DL BWPs containing the SSB associated with the initial DL BWP. For other DL BWPs, the beam-level mobility can only be performed based on CSI-RS. 2.3.4 User Equipment Capability The UE capabilities in NR do not rely on UE categories. Unlike LTE, the NR UE categories are associated with fixed peak data rates and defined for marketing purposes; thus they are not signaled to the network. Instead, the network determines the uplin

k and downlink data rate supported by a UE from the supported band combinations and from the baseband capa- bilities such as modulation scheme, and the number of MIMO layers. In order to limit sig- naling overhead, the gNB can ask the UE to provide NR capabilities for a restricted set of bands. When responding, the UE can skip a subset of the requested band combinations when the corresponding UE capabilities are the same. The NR defines an approximate (peak) data rate for a given number of aggregated carriers in a band or band combination as follows [13]: (Mbps) where J is the number of aggregated component carriers in a band or band combination; Rmax = 948/1024; Valey (j) is the maximum number of supported layers given by higher layer parameter maxNumberMIMO-LayersPDSCH for the downlink and higher layer parameters maxNumberMIMO-LayersCB-PUSCH and maxNumberMIMO-LayersNonCB-PUSCH for the uplink; is the maximum supported modulation order given by higher layer parame- ter supportedModulationOrderDL for the downlink and higher layer parameter supportedModulationOrderUL for the uplink; f(i) is the scaling factor given by higher layer parameter scalingFactor which can take the values of 1, 0.8, 0.75, and 0.4; u is the numer- ology (an OFDM parameter); Ts(u) is the average OFDM symbol duration in a subframe for numerology u which is given as Ts(u) = 0.0142u for normal cyclic prefix; NBWG PRB (u) is the 254 Chapter 2 maximum resource block allocation in bandwidth BW(i) with numerology u where BW(i) is the UE supported maximum bandwidth in the given band or band combination; and (i) is the estimated overhead which takes the values 0.14 [for downlink in frequency range (FR) 1], 0.18 (for downlink in FR2), 0.08 (for uplink in FR1), and 0.10 (for uplink in FR2). Note that only one of the uplink or supplemental uplink carriers with the higher data rate is counted for a cell operating SUL. The approximate maximum data rate can be computed as the maximum of the approximate data rates computed using the above expression for each

of the supported band or band combinations. For LTE in the case of dual connectivity, the approximate data rate for a given number of aggregated carriers in a band or band combination is computed as DMR-DC = 10-3 =1 TBS} (Mbps), where J is the number of aggregated LTE component carriers in multi-radio dual connectivity (MR-DC) band combination and TBSj is the total maximum number of DL-SCH transport block bits received within a 1 ms TTI for the jth component carrier based on the UE supported maximum MIMO layers for the jth carrier, and based on the modulation order and the number of physical resource blocks (PRBs) in the bandwidth of the jth carrier. The approximate maximum data rate can be calculated as the maximum of the approximate data rates computed using the latter equation for each of the supported band or band combinations. For MR-DC, the approximate maximum data rate is computed as the sum of the approximate maximum data rates from NR and LTE [13]. The total layer 2 buffer size is another UE capability attribute that is defined as the sum of the number of bytes that the UE is capable of storing in the RLC transmission windows and RLC reception and reordering windows and also in PDCP reordering windows for all radio bearers. The total layer 2 buffer size in MR-DC and NR-DC scenario is the maximum of the calculated values based on the following equations [13]: MaxULDataRate_MNXRLCRTT_MN+MaxULDataRate_SNXRLCRTT_SN + MaxDLDataRate_SN RLCRTT_SN + MaxDLDataRate_MN X (RLCRTT_SN+X2/Xndelay + QueuinginSN) MaxULDataRate_MNXRLCRTT_MN+ MaxULDataRate_SN X RLCRTT_SN + MaxDLDataRate_MN> RLCRTT_MN + -MaxDLDataRate_SN (RLCRTT_MN + X2/Xndelay+QueuinginMN) In other scenarios, the total layer 2 buffer size is calculated as MaxDLDataRate X RLCRTT + MaxULDataRate X RLCRTT. It must be noted that the additional layer 2 buffer required for preprocessing of data is not taken into account in above formula. The total layer 2 buffer size is determined as the maximum layer 2 buffer size of all calculated ones for each band combinati

on and the applicable Feature Set combination in the supported MR-DC or NR band combinations. The RLC RTT for NR cell group corresponds to the smallest subcarrier spacing (SCS) numerology sup- ported in the band combination and the applicable Feature Set combination. The NR New Radio Access Layer 2/3 Aspects and System Operation 255 specifications specify X2/Xn delay + Queuing in SN = if SCG is NR, and 55 ms if SCG is LTE. The NR specifications define X2/Xn delay + Queuing in MN = 25 ms, if MCG is NR, and 55 ms if MCG is LTE. They further specify RLC RTT for LTE cell group as 75 ms and the RLC RTT for NR cell group ranging from 20 to 50 ms depending on the subcarrier spacing [13]. The maximum supported data rate for integrity-protected DRBs (see Section 2.2.3.3) is a UE capability indicated at NAS layer, with a minimum value of 64 kbps and a maximum value of the highest data rate supported by the UE. In case of failed integrity check (i.e., due to a faulty or missing MAC-I) the corresponding PDU is discarded by the receiving PDCP entity. 2.4 Discontinuous Reception and Power-Saving Schemes The UE monitors physical downlink control channel (PDCCH) while in RRC_CONNECTED state. This activity is controlled by the DRX and bandwidth adaptation schemes configured for the UE. When bandwidth adaptation is configured, the UE only has to monitor PDCCH on the active BWP, that is, it does not have to monitor PDCCH on the entire downlink frequency of the cell. A BWP inactivity timer (independent from the DRX inactivity timer) is used to switch the active BWP to the default one. The latter timer is restarted upon successful PDCCH decoding and the switching to the default BWP happens when it expires. When DRX is configured, the UE is not required to continuously monitor the PDCCH. The DRX mechanism is characterized by the following parameters [8]: On-duration: The time interval during which the UE would expect to receive the PDCCH. If the UE successfully decodes the PDCCH, it stays awake and starts the inac- tivity timer. Inact

ivity timer: The time interval during which the UE waits for successful decoding of the PDCCH, starting from the last successful decoding of a PDCCH. If the decoding fails, the UE can go back to sleep. The UE restarts the inactivity timer following a sin- gle successful decoding of a PDCCH for the first transmission only (i.e., not for retransmissions). Retransmission-timer: The time interval until a retransmission can be expected. Cycle: It specifies the periodic repetition of the on-duration followed by a possible period of inactivity. Active-time: The total time duration that the UE monitors PDCCH. This includes the on-duration of the DRX cycle, the time that the UE is performing continuous reception while the inactivity timer is running, and the time when the UE is performing continu- ous reception while awaiting a retransmission opportunity. 256 Chapter 2 Due to bursty nature of the packet data traffic, which is characterized by intermittent periods of transmission activity followed by longer periods of inactivity, and to reduce the UE power consumption, NR supports a DRX scheme similar to that of LTE. Bandwidth adaptation and dynamic carrier activation/deactivation are two other power- saving mechanisms supported in NR. The underlying mechanism for DRX is a configur- able DRX cycle in the device. When a DRX cycle is configured, the device monitors the downlink control channel only during the active-time and sleeps, with its receiver circuitry switched off, during the inactivity time, leading to a significant reduction in UE power consumption. The longer the DRX inactive time, the lower the power con- sumption. However, this would have certain implications for the scheduler, since the device is only reachable when it is active according to the DRX cycle configured for it. In many cases, if the device has been scheduled and is engaged in receiving or transmit- ting data, it is likely that it will be scheduled again soon; thus waiting until the next activity period according to the DRX cycle would result in ad

ditional delays. Therefore, to reduce the delays, the device remains in the active state for a configurable period of time after being scheduled. This is realized by an inactivity timer started by the UE every time that it is scheduled where the UE remains awake until the time expires (Fig. 2.25). Since NR supports multiple numerologies the time unit of the DRX timers is specified in milliseconds in order to avoid associating the DRX periodicity to a certain numerology (Fig. 2.26). The NR HARQ retransmissions are asynchronous in both downlink and uplink. If the device has been scheduled a transmission in the downlink that it cannot decode, a typical gNB behavior is to retransmit the data at a later time. In practice, the DRX scheme has a config- urable timer which is started after an erroneously received transport block and is used to wake up the UE receiver when it is likely for the gNB to schedule a retransmission. The value of the timer is preferably set to match the (implementation-specific) roundtrip time in the HARQ protocol. The above mechanism is a (long) DRX cycle in conjunction with the device remaining awake for a period of time after being scheduled. However, in some ser- vices such voice over IP, which is characterized by periods of regular transmission, fol- lowed by periods of no activity, a second (short) DRX cycle can be optionally configured in addition to the long DRX cycle. The RRC entity controls the DRX operation by configuring the following parameters [14]: drx-onDurationTimer: The duration at the beginning of a DRX cycle. drx-SlotOffset: The delay before starting the drx-onDurationTimer. drx-InactivityTimer: The duration after the PDCCH occasion in which a PDCCH indi- cates a new uplink/downlink transmission for the MAC entity. drx-RetransmissionTimerDL (per-DL HARQ process except for the broadcast process): The maximum duration until a downlink retransmission is received. Last successful PDCCH decoding Short DRX cycle timer Inactivity On duration On duration On duration On duration On dur

ation timer (PDCCH monitoring occasions) (PDCCH monitoring occasions) (PDCCH monitoring occasions) (PDCCH monitoring occasions) (PDCCH monitoring occasions) Inactivity timer expiry Short DRX Short DRX Long DRX or MAC control cycle cycle Short DRX cycle element reception cycle expiry Condition 1 satisfied Condition 2 satisfied Figure 2.25 Illustration of DRX mechanism 258 Chapter 2 Frequency selection detection Registration Bootup Call establish app loading Connected- RRC_IDLE state mode DRX transmission transmission UE-originated call/data transmission CONNECTED state Power Long DRX cycles Short DRX cycles Figure 2.26 Example UE power consumption when transitioning through various RRC states [20]. drx-RetransmissionTimerUL (per-UL HARQ process): The maximum duration until a grant for uplink retransmission is received. drx-LongCycleStartOffset: The long DRX cycle and drx-StartOffset which define the subframe where the long and short DRX cycle starts. drx-ShortCycle (optional): The short DRX cycle. drx-ShortCycleTimer (optional): The duration in which the UE follows the short DRX cycle. drx-HARQ-RTT-TimerDL (per-DL HARQ process except for the broadcast process): The minimum duration before a downlink assignment for HARQ retransmission is expected by the MAC entity. drx-HARQ-RTT-TimerUL (per-UL HARQ process): The minimum duration before an uplink HARQ retransmission grant is expected by the MAC entity. 2.5 Mobility Management, Handover, and UE Measurements The NR performs load balancing through handover and redirection mechanisms upon RRC release and through use of inter-frequency and inter-RAT absolute priorities as well as inter-frequency Qoffset parameters (see Section 2.5.2). The measurements per- formed by a UE for connected mode mobility are classified into three types, namely, intra-frequency NR, inter-frequency NR, and inter-RAT measurements for LTE. For each measurement type, one or several measurement objects can be defined (a measure- ment object defines the carrier frequency to be monitored). For each me

asurement object, one or several reporting configurations can be defined (a reporting configuration defines the reporting criteria). Three reporting criteria are used: (1) event-triggered reporting, (2) periodic reporting, and (3) event-triggered periodic reporting. The associa- tion between a measurement object and a reporting configuration is created by a New Radio Access Layer 2/3 Aspects and System Operation 259 measurement identity (a measurement identity associates one measurement object and one reporting configuration of the same RAT). By using several measurement identities (one for each measurement object, reporting configuration pair), it is possible to associ- ate several reporting configurations to one measurement object and to associate one reporting configuration to several measurement objects [8]. The measurements identity is used when reporting results of the measurements. Measurement quantities are considered separately for each RAT. Measurement commands are used by NG-RAN to instruct the UE to start, modify, or stop measurements. Handover can be performed within the same RAT and/or CN, or it can involve a change of the RAT and/or CN. Inter-system fallback toward LTE RAN is performed for load balancing when 5GC does not support emergency ser- vices or voice services. Depending on certain criteria such as CN interface availabil- ity, network configuration, and radio conditions, the fallback procedure results in either connected-state mobility (handover procedure) or idle state mobility (redirec- tion) [8]. 2.5.1 Network-Controlled Mobility The mobility of the UEs in RRC_CONNECTED state is controlled by the network (net- work-controlled mobility), which is classified into two types of mobility, namely, cell-level mobility and beam-level mobility. The cell-level mobility requires explicit RRC signaling in order to be triggered, which results in handover. The main steps of the inter-gNB handover signaling procedures are illustrated in Fig. 2.27. The inter-gNB handover comprises the fol- lowing steps

[8]: Source Target 1: Handover request Admission control 2: Handover request acknowledge 3. RRC reconfiguration Switch to new cell 4. RRC reconfiguration complete Figure 2.27 Inter-gNB handover procedure [8]. 260 Chapter 2 1. The source gNB initiates handover and issues a Handover Request over Xn interface. 2. The target gNB performs admission control and provides the RRC configuration as part of the Handover ACK. 3. The source gNB provides the RRC configuration to the UE in the Handover Command message, which includes the cell ID and all information required to access the target cell, SO that the UE can access the target cell without detecting that cell's SI. In some cases, the information required for contention-based and contention-free random-access procedure can be included in the Handover Command message. The access information to the target cell may include beam-specific information. 4. The UE moves the RRC connection to the target gNB and replies with the Handover Complete message. The user data can be sent in step 4, if the grant allows. The handover mechanism triggered by RRC signaling requires the UE to reset the MAC entity and reestablish RLC entity. The RRC-triggered handovers with and without PDCP entity reestablishment are both supported in NR. For DRBs using RLC-AM mode, the PDCP entity can either be reestablished along with a security key change or initiate a data recovery procedure without a key change. For DRBs using RLC-UM mode and for SRBs, the PDCP entity can either be reestablished in conjunction with a security key change or to remain as it is without a key change. Data for- warding, in-sequence delivery, and duplication avoidance at handover can be guaranteed when the target gNB uses the same DRB configuration as the source gNB. Timer-based handover failure procedure is supported in NR where an RRC connection reestablishment procedure is used for recovering from handover failure [8]. The beam-level mobility does not require explicit RRC signaling in order to be triggered. The gNB provides

the UE, via RRC signaling, with measurement configuration containing configurations of SSB/CSI resources and resource sets, as well as trigger states for trigger- ing channel and interference measurements and reports. Beam-level mobility is then man- aged at lower layers through physical layer and MAC sublayer control signaling. The RRC sublayer is not required to know about the beam that is used at any given time. The SSB- based beam-level mobility is based on the SSB associated with the initial DL BWP and can only be configured for the initial DL BWPs and for DL BWPs containing the SSB associ- ated with the initial DL BWP. For other DL BWPs the beam-level mobility can only be per- formed based on CSI-RS measurements [8]. 2.5.1.1 Control-Plane Handover Procedures The intra-NR handover includes the preparation and execution phases of the handover pro- cedure performed without 5GC involvement, that is, preparation messages are directly exchanged between the gNBs. The release of the resources at the source gNB during the handover completion phase is triggered by the target gNB. Fig. 2.28 shows the basic Source Target UPF(s) User data User data 0. Mobility control information provided by AMF 1. Measurement control and reports 2. Handover decision 3. Handover request 4. Admission control 5. Handover request acknowledge 6. RAN handover initiation 7. SN status transfer Detach from the source cell and synchronize with the target cell Deliver buffer data and new data from UPF(s) User data Buffer user data from the source 8. RAN handover completion Userldata Userldata 9. Path switch request 10. Path switch in UPF(s) End marker User data 11. Path switch request acknowledge 12. UE context release Figure 2.28 Intra-AMF/UPF handover in NR 8]. 262 Chapter 2 handover procedure where the AMF and the UPF entities do not change. The processing steps of this handover are as follows [8]: 1. The UE context within the source gNB contains information concerning roaming and access restrictions, which were provided either at connection

establishment or at the last tracking area update. 2. The source gNB configures the UE measurement procedures and the UE reports according to the measurement configuration. 3. The source gNB decides to handover the UE, based on measurement reports and radio resource management (RRM) information. The source gNB issues a Handover Request message to the target gNB passing a trans- parent RRC container with necessary information to prepare for the handover at the target gNB. The information includes target cell ID, KgNB*, C-RNTI of the UE in the source gNB, RRM configuration including UE inactive time, basic AS-configuration including antenna info and DL carrier frequency, the current QoS flow to DRB map- ping rules applied to the UE, the SIB1 from source gNB, the UE capabilities for differ- ent RATs, and PDU Session related information, and can further include the UE reported measurement information including beam-related information. The PDU Session related information includes the slice information (if supported) and QoS flow level QoS profile(s). After issuing a Handover Request, the source gNB should not reconfigure the UE, including performing reflective QoS flow to DRB mapping. Admission control may be performed by the target gNB. Slice-aware admission con- trol is performed, if the slice information is sent to the target gNB. If the PDU sessions are associated with non-supported slices, the target gNB would reject those PDU sessions. The target gNB prepares the handover with L1/L2 and sends the Handover Request Acknowledge to the source gNB, which includes a transparent container to be sent to the UE as an RRC message to perform the handover. 7. The source gNB triggers the Uu handover by sending an RRCReconfiguration message to the UE, containing the information required to access the target cell, that is, the tar- get cell ID, the new C-RNTI, and the target gNB security algorithm IDs for the selected security algorithms. It can also include a set of dedicated RACH resources, the association between RACH reso

urces and SSB(s), the association between RACH resources and UE-specific CSI-RS configuration(s), common RACH resources, and the SI of the target cell. 8. The source gNB sends the SN Status Transfer message to the target gNB. 9. The UE synchronizes to the target cell and completes the RRC handover procedure by sending RRCReconfigurationComplete message to target gNB. New Radio Access Layer 2/3 Aspects and System Operation 263 10. The target gNB sends a Path Switch Request message to the AMF to trigger 5GC to switch the downlink data path toward the target gNB and to establish an NG-C inter- face instance toward the target gNB. 11. 5GC switches the downlink data path toward the target gNB. The UPF sends one or more end-marker packets on the old path to the source gNB per PDU session/tunnel and then can release any user-plane/TNL resources toward the source gNB. 12. The AMF confirms the Path Switch Request message with the Path Switch Request Acknowledge message. 13. Upon reception of the Path Switch Request Acknowledge message from the AMF the target gNB sends the UE Context Release message to inform the source gNB about the success of the handover. The source gNB can then release radio and control-plane- related resources associated to the UE context. Any ongoing data forwarding continue. The RRM configuration can include both beam measurement information (for layer 3 mobil- ity) associated to SSB(s) and CSI-RS(s) for the reported cell(s), if both types of measure- ments are available. Also, if carrier aggregation is configured, the RRM configuration can include the list of best cells on each frequency for which measurement information is avail- able. The RRM measurement information can also include the beam measurement for the listed cells that belong to the target gNB. The common RACH configuration for beams in the target cell is only associated to the SSB(s). The network can have dedicated RACH configurations associated to the SSB(s) and/or have dedicated RACH configurations associated to CSI-RS(s) within a ce

ll. The tar- get gNB can only include one of the following RACH configurations in the handover com- mand to enable the UE to access the target cell: Common RACH configuration Common RACH configuration + dedicated RACH configuration associated with SSB Common RACH configuration + dedicated RACH configuration associated with CSI-RS The dedicated RACH configuration allocates RACH resource(s) along with a quality thresh- old to use them. When dedicated RACH resources are provided, they are prioritized by the UE, and the UE does not switch to contention-based RACH resources as long as the quality threshold of those dedicated resources is met. The order to access the dedicated RACH resources is up to UE implementation. 2.5.1.2 User-Plane Handover Procedures The user-plane aspects of intra-NR handover for the UEs in RRC_CONNECTED state include the following principles to avoid loss of user data during handover [8]. During handover preparation, the user-plane tunnels can be established between the source gNB 264 Chapter 2 and the target gNB. During handover execution, the user data can be forwarded from the source gNB to the target gNB. Packet forwarding should be done in order, as long as the packets are received at the source gNB from the UPF or the source gNB buffer has not been emptied. During handover completion, the target gNB sends a path switch request message to the AMF to inform it that the UE has been granted access and the AMF then triggers path switch related 5GC internal signaling and actual path switch of the source gNB to the target gNB in UPF. The source gNB should continue forwarding data, as long as packets are received at the source gNB from the UPF or the source gNB buffer has not been emptied. For RLC-AM bearers, in-sequence delivery, and duplication avoidance, the PDCP SN is maintained on a per DRB basis and the source gNB informs the target gNB about the next downlink PDCP SN to allocate to a packet which does not have a PDCP sequence number yet, neither from source gNB nor from the UPF. For secur

ity synchronization, the HFN is maintained and the source gNB provides the target gNB one reference HFN for the uplink and one for the downlink, that is, HFN and corresponding SN. In both UE and target gNB, a window-based mechanism is used for duplication detection and reordering. The occur- rence of duplicates over the air interface in the target gNB is minimized by means of PDCP-SN-based reporting at the target gNB by the UE. In the uplink, the reporting is optionally configured on as per DRB basis by the gNB and the UE initially starts by trans- mitting those reports when granted resources are in the target gNB. In the downlink, the gNB can decide when and for which bearers a report is sent, and the UE does not wait for the report to resume UL transmission [8]. The target gNB retransmits and prioritizes all downlink data forwarded by the source gNB excluding the PDCP SDUs for which the reception was acknowledged through PDCP-SN- based reporting by the UE, that is, the target gNB should initially send all forwarded PDCP SDUs with PDCP SNs, then all forwarded downlink PDCP SDUs without SNs before send- ing new data from 5GC. Lossless delivery, when a QoS flow is mapped to a different DRB at handover, requires that the old DRB to be configured in the target cell. For in-order delivery in the downlink, the target gNB should first transmit the forwarded PDCP SDUs on the old DRB before transmitting new data from 5G CN on the new DRB. In the uplink, the target gNB should not deliver data of the QoS flow from the new DRB to 5G CN before receiving the end-marker on the old DRB from the UE [8]. The UE retransmits in the target gNB all uplink PDCP SDUs starting from the oldest PDCP SDU that has not been acknowledged at RLC sublayer in the source, excluding PDCP SDUs for which the reception was acknowledged through PDCP-SN-based reporting by the target. For RLC-UM bearers the PDCP SN and HFN are reset in the target gNB; no PDCP SDUs are retransmitted in the target gNB; and the target gNB prioritizes all downlink SDAP SDUs

forwarded by the source gNB over the data from the CN. To minimize the New Radio Access Layer 2/3 Aspects and System Operation 265 losses when a QoS flow is mapped to a different DRB at handover, the old DRB needs to be configured in the target cell. For in-order delivery in the downlink, the target gNB should first transmit the forwarded PDCP SDUs on the old DRB before transmitting new data from 5G CN on the new DRB. In the uplink, the target gNB should not deliver data of the QoS flow from the new DRB to 5G CN before receiving the end-marker on the old DRB from the UE. The UE does not retransmit any PDCP SDU in the target cell for which transmis- sion had been completed in the source cell. The source NG-RAN node may request downlink data forwarding per QoS flow to be estab- lished for a PDU session and may provide information on how it maps QoS flows to DRBs. The target NG-RAN node decides whether data forwarding per QoS flow should be estab- lished for a PDU session. If lossless handover is desired and the QoS flow to DRB mapping, applied at the target NG-RAN node, allows employing data forwarding with the same QoS flow to DRB mapping that was used in the source NG-RAN node for a DRB and if all QoS flows mapped to that DRB are accepted for data forwarding, the target NG-RAN node establishes a downlink forwarding tunnel for that DRB. For a DRB for which SN status preservation is important, the target NG-RAN node may decide to establish an uplink data forwarding tunnel. The target NG-RAN node may also decide to establish a downlink forwarding tunnel for each PDU session. In this case, the target NG-RAN node provides information related to the QoS flows for which data forwarding has been accepted and the corresponding uplink TNL information for data forwarding tunnels to be established between the source and the target NG-RAN nodes [8]. 2.5.2 UE-Based Mobility The PLMN selection in NR is based on 3GPP PLMN selection rules. Cell selection is required upon transition from RM-DEREGISTERED to RM-REGISTERED, from CM-

IDLE to CM-CONNECTED, and from CM-CONNECTED to CM-IDLE. The UE NAS layer identi- fies a selected PLMN and equivalent PLMNs. Cell selection is always based on CD-SSBs located on the synchronization raster. The UE scans the NR frequency bands and for each car- rier frequency identifies the strongest cell and the associated CD-SSB. It then detects broad- cast SI of the cell to identify its PLMN(s). The UE may scan each carrier in certain order during initial cell selection or take advantage of stored information to expedite the search dur- ing stored information cell selection. The UE then tries to identify a suitable cell. If it is not able to identify a suitable cell, it then tries to identify an acceptable cell. When a suitable cell or an acceptable cell is found, the UE camps on it and begins the cell reselection procedure. A suitable cell is a cell whose measured cell attributes satisfy the cell selection criteria; the cell PLMN is the selected PLMN, registered or an equivalent PLMN; the cell is not barred or 266 Chapter 2 reserved, and the cell is not part of a tracking area which is in the list of forbidden tracking areas for roaming. An acceptable cell is a cell whose measured cell attributes satisfy the cell selection criteria, and the cell is not barred. Upon transition from RRC_CONNECTED or RRC_INACTIVE to RRC_IDLE state, a UE may camp on a cell as a result of cell selection according to the frequency assigned to the UE via RRC signaling in the state transition message. The UE may attempt to find a suitable cell in the above manner described for stored information or initial cell selection. If no suitable cell is found on any frequency or RAT, the UE may attempt to find an acceptable cell. In multi-beam operations, the cell quality is derived among the beams that are correspond- ing to the same cell [8]. The cell selection criterion S is considered fulfilled, if the following criteria are satisfied: Srxlev > 0 AND Squal > 0 where Srxlev = Qrxlevmeas - (Qrxlevmin + Qrxlevminoffset) - P compensation Qoffset

temp and Squal = Qqualmeas - (Qqualmin + Qqualminoffset) - Qoffsettemp. The signaled values Qrxlevminoffset and Qqualminoffset are only applied when a cell is evaluated for cell selection as a result of a periodic search for a higher priority PLMN, while camped nor- mally in a visited PLMN (VPLMN). During the periodic search for higher priority PLMN, the UE may check the S criterion of a cell using stored parameter values that were obtained from a different cell within the higher priority PLMN [12]. The cell selection parameters are described in Table 2.4. The following rules are used by the UE to limit the required measurements. If the serving cell fulfills Srxlev > SIntraSearchP AND Squal > SIntraSearchQ, the UE may skip intra- frequency measurements; otherwise, the UE must perform intra-frequency measurements. The UE must apply the following rules for NR inter-frequency and inter-RAT frequency measurements, which are identified in the SI and for which the UE has priority. For an NR inter-frequency or inter-RAT frequency with a reselection priority higher than the reselection priority of the current NR frequency, the UE is required to perform measure- ments on higher priority NR inter-frequency or inter-RAT frequencies [6]. For an NR inter-frequency with an equal or lower reselection priority than the reselection priority of the current NR frequency and for inter-RAT frequency with lower reselection priority than the reselection priority of the current NR frequency, if the serving cell fulfills Srxlev > SnonIntraSearchP AND Squal > SnonIntraSearchQ criterion, the UE may skip measure- ments of NR inter-frequencies or inter-RAT frequency cells of equal or lower priority; otherwise, the UE is required to perform measurements on NR inter-frequencies or inter- RAT frequency cells of equal or lower priority [6,8]. New Radio Access Layer 2/3 Aspects and System Operation 267 Table 2.4: Cell selection parameters [7,12]. Parameter Description Srxlev Cell selection receive-level value (dB) Squal Cell selection quality val

ue (dB) Qoffsettemp An offset temporarily applied to a cell (dB) Qrxlevmeas Measured cell receive-level value (RSRP) Qqualmeas Measured cell quality value (RSRQ) Qrxlevmin Minimum required receive-level in the cell (dBm). If the UE supports SUL frequency for this cell, Qrxlevmin is obtained from RxLevMinSUL, if present, in SIB1, SIB2, and SIB4. If QrxlevminoffsetcellSUL is present in SIB3 and SIB4 for the candidate cell, this cell-specific offset is added to the corresponding Qrxlevmin to achieve the required minimum receive-level in the candidate cell; otherwise, Qrxlevmin is obtained from q-RxLevMin in SIB1, SIB2, and SIB4. If Qnxlevminoffsetcell is present in SIB3 and SIB4 for the candidate cell, this cell-specific offset is added to the corresponding Qrxlevmin to achieve the required minimum receive-level in the candidate cell Qqualmin Minimum required quality level in the cell (dB). If Qqualminoffsetcell is signaled for the concerned cell, this cell-specific offset is added to achieve the required minimum quality level in the candidate cell Qrxlevminoffset An offset to the signaled Qrxlevmin taken into account in the Srxlev evaluation as a result of periodic search for a higher priority PLMN while camped normally in a VPLMN Qqualminoffset Offset to the signaled Qqualmin taken into account in the Squal evaluation as a result of periodic search for a higher priority PLMN while camped normally in a VPLMN P compensation If the UE supports the additional Pmax in the NR-NS-PmaxList, if present, in SIB1, SIB2, and SIB4, max(PEMAX1 - PPowerClass, 0) - - min(PEMAX1, PPowerClass) (dB) else max(PEMAX1 -PpowerClass,0)(dB) - PEMAX1, PEMAX2 Maximum transmit-power level that a UE may use when transmitting in the uplink in the cell (dBm) is defined as PEMAX. If the UE supports SUL frequency for this cell, PEMAX1 andP EMAX2 are obtained from the p-Max for SUL in SIB1 and NR-NS-PmaxList for SUL, respectively, in SIB1, SIB2, and SIB4; otherwise, PEMAX1 and PEMAX2 are obtained from the p- Max and NR-NS-PmaxList, respectively, i

n SIB1, SIB2, and SIB4 for regular uplink PPowerClass Maximum RF output power of the UE (dBm) according to the UE power class When evaluating Srxlev and Squal of non-serving cells for reselection evaluation purposes, the UE must use the parameters that are provided by the serving cell and for the verification of cell selection criterion, the UE must use the parameters provided by the target cell for cell reselection. The NAS can control the RAT(s) in which the cell selection should be performed, for instance by indicating RAT(s) associated with the selected PLMN, and by maintaining a list of forbidden registration area(s) and a list of equivalent PLMNs. The UE must select a suitable cell based on RRC_IDLE or RRC_INACTIVE state measurements and cell selection criteria. In order to expedite the cell selection process, the previously stored information for other RATs may be used by the UE. After camping on a cell, the UE is required to regularly search for a better cell according to the cell reselection criteria. If a better cell is found, the UE proceed to select that cell. The change of cell may imply a change of RAT. 268 Chapter 2 The UE mobility state is determined, if the parameters and TCRmaxHyst) are broadcast in SI for the serving cell. The state detection criteria are based on the following principles [12]: Normal-mobility state criterion: If the number of cell reselections during time period TCRmax is less than NCR_M. Medium-mobility state criterion: If the number of cell reselections during time period TCRmax is greater than or equal to NCR_M but less than or equal to NCR_M High-mobility state criterion: If the number of cell reselections during time period TCRmax is greater than NCR_M The UE must not attempt to make consecutive reselections, where a cell is reselected again immediately after one reselection for mobility state detection criteria. The state transitions of the UE are determined based on the following criteria [12]: If the criteria for high-mobility state is detected, the UE must transition

to the high-mo- bility state. If the criteria for medium-mobility state is detected, the UE must transition to medium-- mobility state. If criteria for either medium- or high-mobility state is not detected during time period TCRmaxHyst, the UE must transition to normal-mobility state. If the UE is in high- or medium-mobility state, it must apply the speed-dependent scaling rules [12]. 2.5.3 Paging Paging allows the network to reach UEs in RRC_IDLE and RRC_INACTIVE states, and to notify the UEs in RRC_IDLE, RRC_INACTIVE, and RRC_CONNECTED states of SI change, as well as to send ETWS/CMAS notifications. While in RRC_IDLE state, the UE monitors the PCHs for CN-initiated paging. The UEs in RRC_INACTIVE state also monitor PCHs for RAN-initiated paging. To ensure limiting the adverse impact of pag- ing procedure on the battery consumption, a UE does not need to continuously monitor the PCHs since NR defines a UE-specific paging DRX. The UE in RRC_IDLE or RRC_INACTIVE state is only required to monitor PCHs during paging occasions in a DRX cycle. The paging DRX cycles are configured by the network as follows [8]. For CN-initiated paging, a default cycle is broadcast in SI and also a UE-specific cycle is configured via NAS signaling. For RAN-initiated paging, a UE-specific cycle is config- ured via RRC signaling. The UE applies the shortest of the DRX cycles that are config- ured for it, that is, a UE in RRC_IDLE uses the shortest of the CN-initiated paging cycles, whereas a UE in RRC_INACTIVE applies the shortest of CN-initiated and RAN- initiated paging cycles (see Fig. 2.29). New Radio Access Layer 2/3 Aspects and System Operation 269 Paging frame Paging frame in for UE with idle-mode DRX UE_IDx cycle TDRX: DRX cycle Paging occasion Paging occasion for UE with in each paging UE_IDx frame Figure 2.29 Example illustration of NR paging frames and paging occasions [12]. In CN-initiated and RAN-initiated paging, the POs of a UE are based on the same UE_ID, resulting in overlapping POs for both cases. The number of different

POs in a DRX cycle is configurable via SI and a network may distribute UEs into those POs based on their UE_IDs. When in RRC_CONNECTED state, the UE monitors the PCHs in any PO signaled in the SI for SI change indication and public warning system notification. In the case of bandwidth adaptation, a UE in RRC_CONNECTED state only monitors PCHs on the active BWP with common search space configured (see Chapter 4). To optimize the paging procedure for the UEs in CM-IDLE state, upon UE context release, the NG-RAN node may provide the AMF with a list of recommended cells and NG-RAN nodes as assistance information for subsequent paging of the UE. The AMF may further provide paging attempt information consisting of a paging attempt count and the intended number of paging attempts and possibly the next paging area scope. If paging attempt infor- mation is included in the paging message, each paged NG-RAN node receives the same information during a paging attempt. The paging attempt count is increased by one at each new paging attempt. The next paging area scope, when present, indicates whether the AMF intends to modify the paging area currently selected at the next paging attempt. If the UE has changed its state to CM-CONNECTED, the paging attempt count is reset [8]. To optimize the paging procedure for the UEs in RRC_INACTIVE state, upon RAN-initiated paging, the serving NG-RAN node provides the RAN paging area informa- tion. The serving NG-RAN node may also provide the RAN paging attempt information. Each paged NG-RAN node receives the same RAN paging attempt information during a paging attempt with the following content: paging attempt count, the intended number of paging attempts and the next paging area scope. The paging attempt count is increased by 270 Chapter 2 one at each new paging attempt. The next paging area scope, when present, indicates whether the serving NG_RAN node plans to modify the RAN paging area currently selected for the next paging attempt. If the UE leaves RRC_INACTIVE state, the paging attempt

count is reset [8]. When a UE is paged, the paging message is broadcast over a group of cells. In NR the basic principle of UE tracking is the same for idle and inactive modes, although the grouping is to some extent different in the two cases. The NR cells are grouped into RAN areas, where each RAN area is identified by a RAN area ID (RAI). The RAN areas are grouped into larger tracking areas, where each tracking area is identified by a tracking area ID (TAI). As a result, each cell belongs to one RAN area and one tracking area, the identities of which are provided as part of the cell SI. The tracking areas are the basis for device tracking at core-network level. Each device is assigned a UE registration area by the CN, consisting of a list of TAIs. When a device enters a cell that belongs to a tracking area not included in the assigned UE registration area, it accesses the network, including the CN, and per- forms a NAS registration update. The CN registers the device location and updates the device registration area and provides the device with a new TAI list that includes the new TAI. The reason that the device is assigned a set of TAIs is to avoid repeated NAS regis- tration updates, every time that the device crosses the border of two neighbor tracking areas. By keeping the old TAI within the updated UE registration area, no new update is needed, if the device moves back to the old TAI. The RAN area is the basis for device tracking on RAN level. The UEs in the inactive mode can be assigned an RNA that con- sists of either a list of cell identities; a list of RAIs; or a list of TAIs [17]. The procedure for RNA update is similar to the update of the UE registration area. When a device enters a cell that is not included in its RNA, it accesses the network and per- forms an RNA update. The radio network registers the device location and updates the device RNA. Since the change of tracking area always implies the change of the device RAN area, an RNA update is implicitly performed every time a device performs a

UE registration update. In order to track its movement within the network, the device searches for and measures SSBs similar to the initial cell search procedure. Once the device detects an SSB with a received power that exceeds the received power of its current SSB by a certain threshold, it detects the SIB1 of the new cell in order to acquire information about the tracking and RAN areas. 2.5.4 Measurements The UE in RRC_CONNECTED state conducts measurements on multiple beams (at least one) of a cell and averages the measurement results mainly in the form of power values, in order to derive the cell quality. Therefore, the UE is configured to consider a sub- set of the detected beams. Filtering is applied at two different levels namely at the physical New Radio Access Layer 2/3 Aspects and System Operation 271 RRC configures RRC configures RRC configures parameters parameters parameters UE implementation specific gNB beam 1 Layer 1 filtering Evaluation of gNB beam 2 Layer 1 filtering Layer 3 filtering consolidation/ reporting selection Cell quality for cell quality criteria gNB beam K Layer 1 filtering Nbest-beams K beams K beams Beams L3 beam filtering L3 beam filtering Beam selection for reporting L3 beam filtering RRC configures RRC configures parameters parameters Figure 2.30 NR RRM measurement model [8]. layer to derive beam quality and then at RRC level to derive cell quality from multiple beams. The cell quality from beam measurements is derived in the same way for the serving cell(s) and for the non-serving cell(s). The measurement reports may contain the measure- ment results of the Nbest-beams best beams, if the UE is configured by the gNB. The K beams correspond to the measurements on SSB or CSI-RS resources configured for layer 3 mobil- ity by gNB and detected by UE at layer 1. The corresponding high-level measurement model is illustrated in Fig. 2.30 where the para- meters can be described as follows [8]: A: The measurements (beam-specific samples) internal to the physical layer. Layer 1 filtering:

Internal layer 1 filtering of the inputs measured at point A. The exact filtering function is implementation specific and the way that the measurements are con- ducted in the physical layer by an implementation (inputs A and layer 1 filtering) in not specified by the standard. Aj The measurements (i.e., beam-specific measurements) reported by layer 1 to layer 3 after layer 1 filtering. Beam consolidation/selection: The beam-specific measurements are consolidated to derive cell quality. The behavior of the beam consolidation/selection is standardized, and the configuration of this module is provided via RRC signaling. The reporting period at B equals one measurement period at A1. B: A measurement (i.e., cell quality) derived from beam-specific measurements reported to layer 3 after beam consolidation/selection. 272 Chapter 2 Layer 3 filtering for cell quality: The filtering performed on the measurements provided at point B. The behavior of the layer 3 filters is standardized and the configuration of the layer 3 filters is provided via RRC signaling. The filtering reporting period at C equals one measurement period at B. C: A measurement after processing in the layer 3 filter. The reporting rate is identical to the reporting rate at point B. This measurement is used as input for one or more evalua- tion of reporting criteria. Evaluation of reporting criteria: It checks whether actual measurement reporting is nec- essary at point D. The evaluation can be based on more than one flow of measurements at reference point C, for example, to compare different measurements. This is illustrated by inputs C and C1. The UE evaluates the reporting criteria at least every time a new measurement result is reported at points C and C1. The reporting criteria are standard- ized, and the configuration is provided via RRC signaling (UE measurements). D: The measurement report information (message) sent on the radio interface. L3 beam filtering: The filtering performed on the measurements (i.e., beam-specific measurements) provided at

point A 1 The behavior of the beam filters is standardized, and the configuration of the beam filters is provided via RRC signaling. The filtering reporting period at E equals one measurement period at A . E: A measurement (i.e., beam-specific measurement) after processing in the beam filter. The reporting rate is identical to the reporting rate at point A1. This measurement is used as input for selecting the Nbest-beams measurements to be reported. Beam selection for beam reporting: It selects the Nbest-beams measurements from the measurements provided at point E. The behavior of the beam selection is standardized, and the configuration of this module is provided via RRC signaling. F: The beam measurement information included in measurement report sent over the radio interface. Layer 1 filtering introduces a certain level of measurement averaging and the manner through which the UE performs the required measurements is implementation specific to the point that the output at B fulfills the performance requirements. The layer 3 filtering function for cell quality and the related parameters do not introduce any delay in the sample availability at points B and C in Fig. 2.30. The measurements at points C and C1 are the input used in the event evaluation. The L3 beam filtering and the related parameters do not cause any delay in the sample availability at points E and F. The measurement reports are characterized by the following criteria [8]: Measurement reports include the measurement identity of the associated measurement configuration that triggered the reporting. Cell and beam measurement quantities to be included in measurement reports are con- figured by the network. New Radio Access Layer 2/3 Aspects and System Operation 273 The number of non-serving cells to be reported can be limited through configuration by the network. Cells belonging to a blacklist configured by the network are not used in event evalua- tion and reporting, and conversely when a whitelist is configured by the network, only the cells belon

ging to the whitelist are used in event evaluation and reporting. Beam measurements to be included in measurement reports are configured by the net- work (beam ID only, measurement result and beam ID, or no beam reporting). The intra-frequency neighbor (cell) measurements and inter-frequency neighbor (cell) mea- surements are defined as follows [8]: SSB-based intra-frequency measurement: A measurement is defined as an SSB-based intra-frequency measurement provided that the center frequency of the SSB of the serv- ing cell and the center frequency of the SSB of the neighbor cell are the same, and the subcarrier spacing of the two SSBs are also the same. SSB-based inter-frequency measurement: A measurement is defined as an SSB-based inter-frequency measurement provided that the center frequency of the SSB of the serv- ing cell and the center frequency of the SSB of the neighbor cell are different, or the subcarrier spacing of the two SSBs are different. It must be noted that for SSB-based measurements, one measurement object corresponds to one SSB and the UE considers different SSBs as different cells. CSI-RS-based intra-frequency measurement: A measurement is defined as a CSI-RS-based intra-frequency measurement, if the bandwidth of the CSI-RS resource on the neighbor cell configured for measurement is within the bandwidth of the CSI-RS resource on the serving cell configured for measurement, and the subcarrier spacing of the two CSI-RS resources is the same. CSI-RS-based inter-frequency measurement: A measurement is defined as a CSI-RS-based inter-frequency measurement, if the bandwidth of the CSI-RS resource on the neighbor cell configured for measurement is not within the bandwidth of the CSI-RS resource on the serving cell configured for measurement, or the subcarrier spac- ing of the two CSI-RS resources are different. A measurement can be non-gap-assisted or gap-assisted depending on the capability of the UE, the active BWP of the UE and the current operating frequency, described as follows [8]: For an SSB-b

ased inter-frequency measurement, a measurement gap configuration is always provided if the UE only supports per-UE measurement gaps or if the UE sup- ports per-FR measurement gaps and any of the configured BWP frequencies of any of the serving cells are in the same FR of the measurement object. 274 Chapter 2 For an SSB-based intra-frequency measurement, a measurement gap configuration is always provided in the case where, other than the initial BWP, if any of the UE config- ured BWPs do not contain the frequency domain resources of the SSB associated with the initial DL BWP. In non-gap-assisted scenarios, the UE must be able to conduct measurements without mea- surement gaps. In gap-assisted scenarios, the UE cannot be assumed to be able to conduct measurements without measurement gaps. 2.6 UE and Network Identifiers An NR UE in the connected mode uses a number of network-assigned temporary identifiers in order to communicate to gNB and 5GC. Those identifiers, their descriptions, and their usage are summarized as follows [8]: C-RNTI: A unique UE identification used as an identifier of the RRC connection and for scheduling purposes. CS-RNTI: A unique UE identification used for SPS in the downlink or configured grant in the uplink. INT-RNTI: An identification of preemption in the downlink. P-RNTI: An identification of paging and SI change notification in the downlink. SI-RNTI: An identification of broadcast and SI in the downlink. SP-CSI-RNTI: Unique UE identification used for semi-persistent CSI reporting on PUSCH. The following identifiers are used for power and slot format control [8]: SFI-RNTI: An identification of slot format. TPC-PUCCH-RNTI: A unique UE identification to control the power of PUCCH. TPC-PUSCH-RNTI: A unique UE identification to control the power of PUSCH. TPC-SRS-RNTI: A unique UE identification to control the power of SRS. The following identities are used during random-access procedure [8]: RA-RNTI: An identification of the RAR message in the downlink. Temporary C-RNTI: A UE identification

Logical Channel Channel P-RNTI Paging and system information change notification SI-RNTI Broadcast of system information DL-SCH RA-RNTI Random-access response DL-SCH Temporary C-RNTI Contention resolution DL-SCH (when no valid C-RNTI is available) Temporary C-RNTI Msg3 transmission UL-SCH CCCH, DCCH, C-RNTI, MCS-C-RNTI Dynamically scheduled unicast transmission UL-SCH DCCH, DTCH C-RNTI Dynamically scheduled unicast transmission DL-SCH CCCH, DCCH, MCS-C-RNTI Dynamically scheduled unicast transmission DL-SCH DCCH, DTCH C-RNTI Triggering of PDCCH ordered random access CS-RNTI Configured scheduled unicast transmission DL-SCH, DCCH, DTCH (activation, reactivation, and retransmission) UL-SCH CS-RNTI Configured scheduled unicast transmission (deactivation) TPC-PUCCH-RNTI PUCCH power control TPC-PUSCH-RNTI PUSCH power control TPC-SRS-RNTI SRS trigger and power control INT-RNTI Indication of preemption in the downlink SFI-RNTI Slot format indication in a given cell SP-CSI-RNTI Activation of semi-persistent CSI reporting on PUSCH A complete list of UE RNTIs is provided in Table 2.5. The following identities are used in NG-RAN for identifying a specific network entity [8]: AMF name: This is used to identify an AMF. NCGI: This is used to globally identify the NR cells. The NCGI is constructed from the PLMN identity to which the cell belongs to and the NR cell identity (NCI) of the cell. gNB ID: This is used to identify the gNBs within a PLMN. The gNB ID is contained within the NCI of its cells. Global gNB ID: This is used to globally identify the gNBs. The global gNB ID is con- structed from the PLMN identity to which the gNB belongs and the gNB ID. The MCC and MNC are the same as included in the NCGI. Tracking area identity (TAI): This is used to identify tracking areas. The TAI is con- structed from the PLMN identity the tracking area belongs to and the tracking area code of the tracking area. 276 Chapter 2 Single network slice selection assistance information (S-NSSAI): This is used to iden- tify a network slice. 2.7 Rand

om-Access Procedure (L2/L3 Aspects) The random-access procedure is triggered by a number of events including initial access from RRC_IDLE state; RRC connection reestablishment procedure, handover, downlink/ uplink data arrival when in RRC_CONNECTED state and if uplink synchronization status is non-synchronized; uplink data arrival while in RRC_CONNECTED state and when there are no available PUCCH resources for SR; SR failure; request by RRC upon synchronous reconfiguration; transition from RRC_INACTIVE state; establishing timing alignment upon SCell addition; request for other SI; and beam failure recovery. Furthermore, the random- access procedure takes two distinct forms: contention-based random access and contention- free random access as shown in Fig. 2.31. For random access in a cell configured with SUL, the network can explicitly signal which carrier to use (uplink or SUL); otherwise, the UE selects the SUL carrier, if the measured quality of the downlink is lower than a broad- cast threshold. Once started, all uplink transmissions of the random-access procedure remain on the selected carrier [8]. The complete description of the random-access procedure includ- ing the physical layer aspects can be found in Chapter 3. Random access preamble Random access preamble assignment Random access response Random access preamble Scheduled transmission Random access response Contention resolution Contention-based random access procedure Contention-free random access procedure Figure 2.31 Contention-based and contention-free random-access procedures [8]. New Radio Access Layer 2/3 Aspects and System Operation 277 2.8 Multi-radio Dual Connectivity (L2/L3 Aspects) MR-DC is a generalization of the intra-E-UTRA dual connectivity. The NG-RAN supports MR-DC operation wherein a UE in RRC_CONNECTED state is configured to utilize radio resources provided by (at least) two distinct schedulers, located in two different NG-RAN nodes connected via a non-ideal backhaul, one providing the NR access and the other pro- viding either E-U

TRA or NR access. In MR-DC, one network node acts as the master node (MN) and the other as the secondary node (SN). The MN and SN entities are connected via a network interface and the MN is connected to the CN. The NR MR-DC scheme is designed based on the assumption of non-ideal backhaul between different nodes but can also be used in the case of ideal backhaul. The LTE network supports MR-DC via EN-DC, in which a UE is connected to one eNB that acts as the MN and one en-gNB 11 that acts as a SN. The eNB is connected to the EPC via the S1 interface and to the en-gNB via the X2 interface. The en-gNB may also be connected to the EPC via the S1-U interface and other en-gNBs via the X2-U interface [4]. The NG-RAN supports NG-RAN EN-DC (NGEN-DC), in which a UE is connected to one ng-eNB that acts as the MN and one gNB that acts as an SN. The ng-eNB is connected to the 5GC and the gNB is connected to the ng-eNB via the Xn interface. The NG-RAN fur- ther supports NR-E-UTRA DC (NE-DC), in which a UE is connected to one gNB that acts as the MN and one ng-eNB that acts as an SN. The gNB is connected to 5GC and the ng- eNB is connected to the gNB via the Xn interface. Another important scenario is NR-NR dual connectivity (NR-DC), in which a UE is connected to one gNB that acts as the MN and another gNB that acts as the SN. The master gNB is connected to the 5GC via the NG interface and to the secondary gNB via the Xn interface. The secondary gNB may also be connected to the 5GC via the NG-U interface. In addition, NR-DC can also be used when a UE is connected to two gNB-DUs, one serving the MCG and the other serving the SCG, connected to the same gNB-CU, acting both as the MN and the SN. When the UE is config- ured with SCG, it is configured with two MAC entities: one MAC entity for the MCG and one MAC entity for the SCG [4]. In MR-DC, the UE has a single RRC state, based on the MN RRC and a single control- plane connection toward the CN. Each radio node has its own RRC entity (LTE version, if the node is an eNB or NR vers

ion if the node is a gNB) which can generate RRC PDUs to be sent to the UE. The RRC PDUs generated by the SN can be transported via the MN to the UE. The MN always sends the initial SN RRC configuration via MCG SRB (SRB1); however, subsequent reconfigurations may be transported via MN or SN. When en-gNB is a node providing NR user-plane and control-plane protocol terminations toward the UE and acts as a secondary node in EN-DC. A secondary node in MR-DC is a radio access node, with no control-plane connection to the core network, providing only additional radio resources to the UE. It may be an en-gNB (in EN-DC), a secondary ng-eNB (in NE-DC), or a secondary gNB (in NR-DC and NGEN-DC) [4]. 278 Chapter 2 transporting RRC PDU from the SN, the MN does not modify the UE configuration pro- vided by the SN [4]. When an LTE node is connected to the EPC, upon initial connection establishment, SRB1 uses LTE PDCP. If the UE supports EN-DC, regardless of whether EN-DC is configured, after initial connection establishment, the MCG SRBs (SRB1 and SRB2) can be configured by the network to use either LTE PDCP or NR PDCP (either SRB1 and SRB2 are both con- figured with LTE PDCP, or they are both configured with NR PDCP). A change from LTE PDCP to NR PDCP (or vice versa) is supported via a handover procedure (reconfiguration with mobility) or, for the initial change of SRB1 from LTE PDCP to NR PDCP, with a reconfiguration without mobility before the initial security activation. If the SN is a gNB (i.e., the case for EN-DC, NGEN-DC, and NR-DC scenario), the UE can be configured to establish an SRB with the SN (SRB3) to enable RRC PDUs for the SN to be directly transferred between the UE and the SN. The RRC PDUs for the SN can only be transported directly to the UE for SN RRC reconfiguration without any coordination with the MN. Measurement reporting for mobility within the SN can be sent directly from the UE to the SN, if SRB3 is configured. The split SRB is supported for all MR-DC options, allowing duplication of RRC PDUs generat

ed by the MN via the direct path and via the SN. The split SRB utilizes the NR PDCP. The NR Rel-15 specifications do not sup- port duplication of RRC PDUs generated by the SN via the MN and SN paths. In EN-DC, the SCG configuration is maintained in the UE during suspension. The UE releases the SCG configuration (but not the radio bearer configuration) during resumption initiation (see Fig. 2.24). In MR-DC with 5GC, the UE stores the PDCP/SDAP configuration when mov- ing to RRC_INACTIVE state, but it releases the SCG configuration [4]. There are three bearer types in MR-DC from a UE perspective: MCG bearer, SCG bearer, and split bearer. For EN-DC, the network can configure either LTE PDCP or NR PDCP for the MN-terminated MCG bearers, while NR PDCP is always used for all other bearers. In MR-DC with 5GC, NR PDCP is always used for all bearer types. In NGEN-DC, LTE RLC/ MAC is used in the MN, while NR RLC/MAC is used in the SN. In EN-DC, NR RLC/MAC is used in the MN while LTE RLC/MAC is used in the SN. In NR-DC, NR RLC/MAC is used in both MN and SN. From the network perspective, each bearer (MCG, SCG, and split bearer) can be terminated either in MN or in SN. If only SCG bearers are configured for UE, for SRB1 and SRB2, the logical channels are always configured at least in the MCG, that is, this is still an MR-DC configuration and a PCell always exists. If only MCG bearers are configured for a UE, that is, there is no SCG, this is still considered an MR-DC configura- tion, if at least one of the bearers is terminated in the SN [4]. In MR-DC, two or more component carriers may be aggregated over two cell groups. A UE may simultaneously receive or transmit on multiple component carriers depending on its capabilities. The maximum number of configured component carriers for a UE is 32 for New Radio Access Layer 2/3 Aspects and System Operation 279 downlink and uplink. Depending on UE's capabilities, up to 31 component carriers can be configured for an LTE cell group when the NR cell group is configured. For the NR cell

group, the maximum number of configured component carriers for a UE is 16 for downlink and 16 for uplink. A gNB may configure the same physical cell IDs (PCIs) for several NR cells that it serves. To avoid PCI confusion for MR-DC, the NR PCIs may be allocated in a way that an NR cell is uniquely identifiable by a PCell ID. This PCell is in the coverage area of an NR cell included in the MR-DC operation. In addition, the NR PCIs may only be reused in NR cells on the same SSB frequency sufficiently apart from each other. An X2-C/ Xn-C signaling can be used to help identify NR PCIs by including the cell global identifier (CGI) of the PCell in the respective X2AP/XnAP messages and by providing neighbor cell relationship via non-UE-associated signaling [4]. In MR-DC, the UE is configured with two MAC entities: one MAC entity for the MCG and one MAC entity for the SCG. In MR-DC, SPS resources can be configured on both PCell and PSCell. In MR-DC, the BSR configuration, triggering, and reporting are independently performed per cell group. For split bearers, the PDCP data is considered in BSR in the cell group(s) configured by RRC signaling. In EN-DC, separate DRX configurations are pro- vided for MCG and SCG. Both RLC-AM and RLC-UM can be configured in MR-DC for all bearer types (i.e., MCG, SCG, and split bearers). In EN-DC, packet duplication can be applied to carrier-aggregation in the MN and in the SN; however, MCG bearer carrier- aggregation packet duplication can be configured only in combination with LTE PDCP; and MCG DRB carrier-aggregation duplication can be configured only if dual-connectivity packet duplication is not configured for any split DRB. In NGEN-DC, carrier-aggregation packet duplication can only be configured for SCG bearer. In NE-DC, carrier-aggregation packet duplication can only be configured for MCG bearer. In NR-DC, carrier-aggregation packet duplication can be configured for both MCG and SCG bearers. In EN-DC, ROHC can be con- figured for all bearer types. In MR-DC with 5GC, the network may ho

st up to two SDAP pro- tocol entities for each PDU session, one for MN and the other one for SN. The UE is configured with one SDAP protocol entity per PDU session [4]. In MR-DC, the SN is not required to broadcast SI other than for radio frame timing and SFN. The SI for initial configuration is provided to the UE by dedicated RRC signaling via the MN. The UE acquires radio frame timing and SFN of SCG from the LTE primary and secondary synchronization signals and MIB (if the SN is an eNB) and from the NR primary and secondary synchronization signals and MIB (if the SN is a gNB) of the PSCell. Moreover, upon change of the relevant SI of a configured SCell, the network releases and subsequently adds the corresponding SCell (with updated SI), via one or more RRC recon- figuration messages sent on SRB1 or SRB3. If the measurement is configured for the UE in preparation for the secondary node addition procedure, the MN may configure the measure- ment for the UE. In the case of the intra-secondary node mobility, the SN may configure the measurement for the UE in coordination with the MN [4]. 280 Chapter 2 The secondary node change procedure can be triggered by both MN (only for inter- frequency secondary node change) and SN. For secondary node changes triggered by the SN, the RRM measurement configuration is maintained by the SN which also processes the measurement reporting, without providing the measurement results to the MN. Measurements can be configured independently by the MN and by the SN (intra-RAT mea- surements on serving and non-serving frequencies). The MN indicates the maximum num- ber of frequency layers and measurement identities that can be used in the SN to ensure that UE capabilities are not exceeded. If MN and SN both configure measurements on the same carrier frequency then those configurations must be consistent. Each node (MN or SN) can independently configure a threshold for the SpCell quality. When the PCell quality is above the threshold configured by the MN, the UE is still required to per- fo

rm inter-RAT measurements configured by the MN on the SN RAT. When SpCell quality is above the threshold configured by the SN, the UE is not required to perform measure- ments configured by the SN [4]. The measurement reports, configured by the SN, are sent on SRB1 when SRB3 is not con- figured; otherwise, the measurement reports are sent over SRB3. The measurement results related to the target SN can be provided by MN to target SN at MN-initiated SN change procedure. The measurement results of target SN can be forwarded from the source SN to the target SN via MN at SN-initiated SN change procedure. The measurement results corre- sponding to the target SN can be provided by the source MN to the target MN at inter-MN handover with/without SN change procedure [4]. Per-UE or per-FR measurement gaps can be configured, depending on UE capability to sup- port independent FR measurement and network preference. Per-UE gap applies to both FR1 (LTE and NR) and FR2 (NR) bands. For per-FR gap, two independent gap patterns (i.e., FR1 gap and FR2 gap) are configured for FR1 and FR2. The UE may also be configured with a per-UE gap sharing configuration (applying to per-UE gap) or with two separate gap sharing configurations (applying to FR1 and FR2 measurement gaps, respectively). A mea- surement gap configuration is always provided in the following scenarios: for UEs config- ured with LTE inter-frequency measurements; and for UEs that support either per-UE or per-FR gaps, when the conditions to measure SSB-based inter-frequency measurement or SSB-based intra-frequency measurement are satisfied [4]. 2.9 Carrier Aggregation (L2/L3 Aspects) Multiple NR component carriers can be aggregated and simultaneously transmitted to a UE in the downlink or from a UE in the uplink, allowing an increased operating bandwidth and cor- respondingly higher link data rates. The component carriers do not need to be contiguous in New Radio Access Layer 2/3 Aspects and System Operation 281 the frequency domain and can be in the same frequency band or

different frequency bands, resulting in three scenarios: intra-band carrier aggregation with frequency-contiguous compo- nent carriers, intra-band carrier aggregation with non-contiguous component carriers, and inter-band carrier aggregation with non-contiguous component carriers. While the system-level operation for the three scenarios is the same, the architecture and complexity of RF transceivers can be very different. The NR supports up to 16 downlink/uplink carriers of different bandwidths and different duplex schemes, with the minimum and maxi- mum contiguous bandwidth of 5 and 400 MHz per component carrier, respectively [5,21]. A UE capable of carrier aggregation may receive or transmit simultaneously on multiple com- ponent carriers, while a device not capable of carrier aggregation can access one of the com- ponent carriers at any given time. In the case of inter-band carrier aggregation of multiple half-duplex TDD carriers (supplemental uplink or downlink), the transmission direction of dif- ferent carriers does not necessarily have to be the same, which implies that a carrier- aggregation-capable TDD device may need a front-end duplexer, unlike a typical carrier- aggregation-incapable TDD device that does not include a duplexer. In the LTE and NR spe- cifications, the carrier aggregation is treated as a cell, that is, a carrier-aggregation-capable UE is said to able transmit/receive to/from multiple cells. One of these cells is referred to as the PCell that is the cell which the device initially selects and connects to. Once connected to the gNB, one or more SCells can be configured. The SCells can be activated or deceived to meet various application requirements. Different UEs may have different designated cells as their PCell, meaning that the configuration of the PCell is UE-specific. Furthermore, the num- ber of carriers (or cells) does not have to be the same in uplink and downlink. In a typical scenario, there are more downlink carriers than uplink carriers, since there is often more traf- fic i

n the downlink than in the uplink. Furthermore, the RF implementation complexity and cost of operating multiple simultaneously active uplink carriers are often higher than the cor- responding complexity/cost of the downlink. The scheduling grants and radio resource assign- ments can be transmitted on either the same cell as the corresponding data, referred to as self- scheduling, or on a different cell than the corresponding data, referred to as cross-carrier scheduling. The NR carrier aggregation uses L1/L2 control signaling for scheduling the UE in the downlink, and uplink control signaling to transmit HARQ-ACKs. The uplink feedback is typically transmitted on the PCell to allow asymmetric carrier aggregation. In certain use cases where there are a large number of downlink component carriers and a single uplink component carrier, the uplink carrier would be overloaded with a large number of feedback information. To avoid overloading a single carrier, it is possible to configure two PUCCH groups where the feedback corresponding to the first group is transmitted in the uplink of the PCell and the feedback corresponding to the other group of carriers is transmitted on the PSCell. If carrier aggregation is enabled, the UE may receive and transmit on multiple 282 Chapter 2 carriers, but operating multiple carriers is only needed for high data rates, thus is advanta- geous to deactivate unused carriers. Activation and deactivation of component carriers can be done through MAC CEs (see Section 2.2.1), where a bitmap is used to indicate whether a configured SCell should be activated or deactivated. As we mentioned earlier, to ensure reasonable UE power consumption when carrier aggre- gation is configured, an activation/deactivation mechanism of cells is supported. When an SCell is deactivated, the UE no longer needs to receive the corresponding PDCCH or PDSCH, it cannot transmit in the corresponding uplink, it is not required to perform CQI measurements on that cell. On the other hand, when an SCell is activated, the U

E receives PDSCH and PDCCH (if the UE is configured to monitor PDCCH on this SCell) and is expected to be able to perform CQI measurements on that cell. The NG-RAN ensures that while PUCCH SCell (i.e., an SCell configured with PUCCH) is deactivated, SCells of the secondary PUCCH group (i.e., a group of SCells whose PUCCH signaling is associated with the PUCCH on the PUCCH SCell) are activated. The NG-RAN further ensures that SCells mapped to PUCCH SCell are deactivated before the PUCCH SCell is changed or removed. When reconfiguring the set of serving cells, SCells added to the set are initially deactivated and SCells which remain in the set (either unchanged or reconfigured) do not change their activation status. During handover, the SCells are deactivated. When band- width adaptation is configured, only one UL BWP for each uplink carrier and one DL BWP or only one DL/UL BWP pair can be active at any given time in an active serving cell, all other BWPs that the UE is configured with will be deactivated. The UE does not monitor the PDCCH and does not transmit on PUCCH, PRACH, and UL-SCH of the deactivated BWPs [8]. The SCell activation/deactivation is an efficient mechanism to reduce UE power consump- tion in addition to DRX. On a deactivated SCell, the UE neither receives downlink signals nor transmits any uplink signal. The UE is also not required to perform measurements on a deactivated SCell. Deactivated SCells can be used as pathloss reference for measurements in uplink power control. It is assumed that these measurements would be less frequent while the SCell is deactivated in order to conserve the UE power. On the other hand, for an acti- vated SCell, the UE performs normal activities for downlink reception and uplink transmis- sion. Activation and deactivation of SCells is controlled by the gNB. As shown in Fig. 2.32, the SCell activation/deactivation is performed when the gNB sends an activation/deactiva- tion command in the form of a MAC CE. A timer may also be used for automatic deactiva- tion, if no d

ata or PDCCH messages are received on a SCell for a certain period of time. This is the only case in which deactivation can be executed autonomously by the UE. Serving cell activation/deactivation is performed independently for each SCell, allowing the UE to be activated only on a particular set of SCells. Activation/deactivation is not applica- ble to the PCell because it is required to always remain activated when the UE has an RRC connection to the network [8]. New Radio Access Layer 2/3 Aspects and System Operation 283 PCell Data transmission Data transmission MAC activation SCell 1 command Data transmission Data transmission deactivation control ellement command control element Short DRX Long DRX SCell 2 Data transmission Deactivation timer Figure 2.32 Illustration of SCell activation/deactivation procedure [9]. As already mentioned, if the UE is configured with one or more SCells, the network may activate and deactivate the configured SCells. The PCell is always activated. The network activates and deactivates the SCell(s) by sending an activation/deactivation MAC CE described in Section 2.2.1. Furthermore, the UE maintains a sCellDeactivationTim timer per configured SCell (except the SCell configured with PUCCH) and deactivates the associ- ated SCell upon its expiration. The same initial timer value is applied to each instance of the sCellDeactivationTimer and it is configured by RRC signaling. The configured SCells are initially deactivated upon addition and after a handover. The HARQ feedback for the MAC PDU containing SCell activation/deactivation MAC CE is not impacted by PCell, PSCell, and PUCCH SCell interruptions due to SCell activation/deactivation [9]. References 3GPP Specifications12 [1] 3GPP TS 23.501, System Architecture for the 5G System (Release 15), December 2018. [2] 3GPP TS 33.501, Security Architecture and Procedures for 5G System (Release 15), December 2018. [3] 3GPP TS 37.324, NR, Service Data Adaptation Protocol (SDAP) Specification (Release 15), September 2018. [4] 3GPP TS 37.340, Mul

ti-connectivity, Stage 2 (Release 15), December 2018. [5] 3GPP TS 38.104, NR, Base Station (BS) Radio Transmission and Reception (Release 15), December 2018. [6] 3GPP TS 38.133, NR, Requirements for Support of Radio Resource Management (Release 15), December 2018. [7] 3GPP TS 38.215, NR, Physical Layer Measurements (Release 15), March 2018. [8] 3GPP TS 38.300, NR, Overall Description, Stage-2 (Release 15), December 2018. [9] 3GPP TS 38.321, NR, Medium Access Control (MAC) Protocol Specification (Release 15), December 2018. 3GPP TS 38.322, NR, Radio Link Control (RLC) Protocol Specification (Release 15), December 2018. 3GPP specifications can be accessed at http://www.3gpp.org/ftp/Specs/archive/ Chapter 2 [11] 3GPP TS 38.323, NR, Packet Data Convergence Protocol (PDCP) Specification (Release 15), December 2018. [12] 3GPP TS 38.304, NR, User Equipment (UE) Procedures in Idle Mode and RRC Inactive State (Release 15), December 2018. [13] 3GPP TS 38.306, NR, User Equipment (UE) Radio Access Capabilities (Release 15), December 2018. [14] 3GPP TS 38.331, NR, Radio Resource Control (RRC); Protocol Specification (Release 15), December 2018. [15] 3GPP TS 38.401, NG-RAN, Architecture Description (Release 15), December 2018. IETF Specifications13 [16] IETF RFC 5795, The RObust Header Compression (ROHC) Framework, March 2010. Articles, Books, White Papers, and Application Notes [17] E. Dahlman, S. Parkvall, 5G NR: The Next Generation Wireless Access Technology, Academic Press, August 2018. S. Ahmadi, LTE-Advanced: A Practical Systems Approach to Understanding 3GPP LTE Releases 10 and 11 Radio Access Technologies, Academic Press, November 2013. [19] 3GPP RWS-180010, NR Radio Interface Protocols, Workshop on 3GPP Submission Towards IMT-2020, Brussels, Belgium, October 2018. [20] 5G New Radio, ShareTechNote. <http://www.sharetechnote.com> [21] MediaTek, A New Era for Enhanced Mobile Broadband, White Paper, March 2018. IETF specifications can be accessed at https://datatracker.ietf.org/ CHAPTER 3 New Radio Access Physical Layer A

spects (Part 1) This chapter describes the theoretical and practical aspects of physical layer protocols and functional processing in 3GPP new radio. As shown in Fig. 3.1, the physical layer is the lowest protocol layer in baseband signal processing that interfaces with the digital and the analog radio frontends and the physical media (in this case air interface) through which the signal is transmitted and received. The physical layer further interfaces with the medium access control (MAC) sublayer and receives MAC PDUs and processes the transport blocks through channel coding, rate matching, interleaving/scrambling, baseband modulation, layer mapping for multi-antenna transmission, digital precoding, resource element mapping, orthogonal frequency division multiplexing (OFDM) modulation, and antenna mapping. The choice of appropriate modulation and coding scheme as well as multi-antenna transmission mode is critical to achieve the desired reliability/robustness (coverage) and system/user throughput in mobile communications. Typical mobile radio channels tend to be dispersive and time variant and exhibit severe Doppler effects, multipath delay variation, intra-cell and inter-cell interference, and fading. A good and robust design of the physical layer ensures that the system can robustly operate and overcome the above deleterious effects and can provide the maximum throughput and lowest latency under various operating conditions. Chapters 3 and 4 on physical layer in this book are dedicated to systematic design of physical layer protocols and functional blocks of 5G systems, the theoretical background on physical layer procedures, and perfor- mance evaluation of physical layer components. The theoretical background is provided to make the chapter self-contained and to ensure that the reader understands the underlying theory governing the operation of various functional blocks and procedures. While the focus is mainly on the techniques that were incorporated in the design of 3GPP NR physical layer, the author has a

ttempted to take a more generic and systematic approach to the design of physical layer for the IMT-2020 wireless systems SO that the reader can understand and apply the learnings to the design and implementation of any OFDM-based physical layer irrespective of the radio access technology. 5G NR. DOI: https://doi.org/10.1016/B978-0-08-102267-2.00003-8 © 2019 Elsevier Inc. All rights reserved. Chapter 3 NR protocol structure Non access stratum OSI seven-layer network (NAS) Internet protocol (IP) model Application layer Radio resource control (RRC) Presentation layer QoS Flows Session layer Service data adaptation Control/ protocol (SDAP) configuration Transport layer Radio bearers RRC PDUs Packet data convergence Network layer protocol (PDCP) Control/configuration Data-link layer RLC channels Radio link control (RLC) Physical layer Control/configuration Logical channels Medium access control (MAC) Control/configuration Transport channels Physical layer (PHY) Control/configuration/ physical layer measurements Physical channels Figure 3.1 The physical layer in NR protocol stack [19]. In this chapter, we start with the study of the fundamental concepts and common fea- tures/functions in downlink and uplink of the new radio, which include review of the characteristics of wireless channels in sub-6 GHz and mmWave frequency regions as well as the analysis of two-dimensional (2D) and three-dimensional (3D) channel mod- els and propagation effects. We will then begin our top-down approach to physical layer protocols starting with waveforms, orthogonal and non-orthogonal multiple- access schemes, and duplex schemes as well as the operating frequencies of the new radio. The frame structure, OFDM numerologies, time-frequency resources, and resource allocation techniques will be discussed and analyzed from theoretical and practical point of views. New Radio Access Physical Layer Aspects (Part 1) 287 3.1 Channel Models and Propagation Characteristics 3.1.1 Characteristics of Wireless Channels In a wireless communication system

, a signal can travel from the transmitter to the receiver over multiple paths. This phenomenon is referred to as multipath propagation where signal attenuation varies on different paths. This effect also known as multipath fading can cause stochastic fluctuations in the received signal's magnitude, phase, and angle of arrival (AoA). The propagation over different paths is caused by scattering, reflection, diffraction, and refraction of the radio waves by static and moving objects as well as the transmission medium. It is obvious that different propagation mechanisms result in different channel and path loss models. As a result of wave propagation over multipath fading channels, the radio signal is attenuated due to mean path loss as well as macroscopic and microscopic fading. A detailed channel model for the frequency range (FR) from 6 to 100 GHz was developed in 3GPP [24]. It is applicable to bandwidth up to 10% of the carrier frequency (with a limit of 2 GHz), which accounts for the mobility of one of the two terminals (i.e., in a typical cel- lular network, the base station is fixed and the user terminal moves). It further provides sev- eral optional features that can be plugged in to the basic model, in order to simulate spatial consistency (i.e., the radio environment conditions of nearby users are correlated), blockage, and oxygen absorption. This model supports different IMT-2020 test environments includ- ing urban microcell (UMi), urban macrocell (UMa), rural macrocell (RMa), indoor hotspot (InH), and outdoor to indoor, which must be chosen when setting the simulation parameters [4,5,8,24,25]. 3.1.1.1 Path Loss Models In ideal free-space propagation model, the attenuation of RF signal energy between the transmitter and receiver follows inverse-square law. The received power expressed as a function of the transmitted power is attenuated proportional to the inverse of Ls(d), which is called free-space path loss. When the receiving antenna is isotropic, the received signal power can be expressed as follows

[32,33]: where PTX and PRX denote the transmitted and the received signal power, GTX and GRX denote the transmitting and receiving antenna gains, d is the distance between the transmit- ter and the receiver, and l is the wavelength of the RF signal. In mmWave bands, the smal- ler wavelength translates into smaller captured energy at the receive antenna. For example, in the frequency range 3-30 GHz, an additional 20 dB path loss is added due to effective Chapter 3 mmWave effective aperture Microwave effective aperture Transmitter antenna (isotropic radiator) 4t Grx Transmitter antenna Receiver antenna Receive spectral Effective receiver density aperture No =kT.B = Microwave noise bandwidth mm) Wave noise bandwidth Figure 3.2 Definition of gain and aperture in mmWave [32]. aperture reduction. On the other hand, larger bandwidths in mmWave mean higher noise power and lower signal-to-noise ratios (SNRs), for example, from 50 to 500 MHz band- width, the noise power is increased by 10 dB (see Fig. 3.2). Free-space conditions require a direct line of sight (LoS) between the two antennas involved. Consequently, no obstacles must exist in the path between the antennas at both ends. Furthermore, in order to avoid the majority of effects caused by superposition of direct and reflected signals, it is necessary that the first Fresnel zone 1 is completely free of obstacles. The first Fresnel ellipsoid is defined as a rotational ellipsoid with the two anten- nas at its focal points. Within this ellipsoid, the phase difference between two potential paths is less than half a wavelength. The radius b at the center of the ellipsoid can be calcu- lated as b = 8.66 /d/f where b is the radius in meters, d is the distance between the receiver and the transmitter in kilometers, and f is the frequency in GHz [26,27,32]. A Fresnel zone is one of a series of confocal prolate ellipsoidal regions of space between and around a trans- mitting antenna and a receiving antenna system. The regions are used to understand and compute the strength of

waves (such as sound or radio waves) propagating between a transmitter and a receiver, as well as to pre- dict whether obstructions near the line joining the transmitter and receiver will cause significant interference. New Radio Access Physical Layer Aspects (Part 1) 289 It must be noted that all antenna properties and attributes that we consider in this section assumes far-field patterns. The distance from an antenna, where far-field conditions are met, depends on the dimensions of the antenna relative to the wavelength. For smaller antennas, the wave fronts radiated from the antenna become almost parallel at much closer distance compared to electrically large antennas. A good approximation for small antennas is that far-field conditions are reached at r 2 21. For larger antennas, that is, reflector antennas or array antennas where the dimensions of the antenna L are significantly larger compared to the wavelength L, the far-field distance is approximated as r~2L2/A[26,33]. Macroscopic fading is caused by shadowing effects of buildings and natural obstructions and is modeled by the local mean of a fast fading signal. The mean path loss Lp(d) as a function of distance d between the transmitter and receiver is proportional to an nth power of d relative to a reference distance do. In logarithmic scale, it can be expressed as follows: The reference distance do corresponds to a point located in the far-field of the antenna typi- cally 1 km for large cells, 100 m for microcells, and 1 m for indoor channels. In the above equation, Lp(d) is the mean path loss which is typically 10n dB per decade attenuation for d >> do. The value of n depends on the frequency, antenna heights, and propagation environ- ment that is equal to 2 in free space. The studies show that the path loss Lp(d) is a random variable with log-normal distribution about the mean path loss Lp(d). Let X ~ N(0, o2) denote a zero-mean Gaussian random variable with standard deviation o when measured in decibels, then The value of X is often derived empirica

lly based on measurements. A typical value for o is 8 dB. The parameters that statistically describe path loss due to large-scale fading (macro- scopic fading) for an arbitrary location with a specific transmitting-receiving antenna sepa- ration include the reference distance do, the path loss exponent n, and the standard deviation o of X [33]. Microscopic fading refers to the rapid fluctuations of the received signal in time and frequency and is caused by scattering objects between the transmitting and receiving antennas. When the received RF signal is a superposition of independent scattered compo- nents plus an LoS component, the envelope of the received signal r(t) has a Rician Probability Distribution Function (PDF) that is referred to as Rician fading. As the 290 Chapter 3 magnitude of the LoS component approaches zero, the Rician PDF approaches a Rayleigh PDF. Thus where K and Io(r) denote the Rician factor and zero-order modified Bessel function of the first kind. 2 In the absence of LoS path (K = 0), the Rician PDF reduces to Rayleigh distribution. One of the challenges of mobile communications in the higher frequency bands for outdoor access has been to overcome the difficulties in highly varying propagation conditions. Understanding the propagation conditions will be critical to designing an appropriate air interface and determining the type of hardware (particularly the array size) needed for reli- able communications. Extensive measurements over a wide range of frequencies were per- formed by a large number of academic institutions and the industry. Since maintaining link budgets at higher frequencies are challenging, there are few measurements at larger dis- tances for 5G deployment scenarios of interest. Based on the results of the measurements, some important observations were made that helped development of the new channel mod- els. The most notable signal degradation is due to the higher path loss of the bands above 6 GHz relative to sub-6 GHz bands. The additional path loss as result of increas

ing fre- quency need to be compensated by some means such as larger antenna array sizes with higher array gains and MIMO schemes to ensure sufficient and robust links. Due to large variation of propagation characteristics in bands above 6 GHz, the propagation characteris- tics of different frequency bands were independently investigated. The combined effects of all contributors to propagation loss can be expressed via the path loss exponent. In UMi test environment, the LoS path loss in the bands of interest appears to closely follow the free- space path loss model. In lower bands, a higher path loss exponent was observed in NLoS scenarios. The shadow fading in the measurements appears to be similar to lower frequency bands, while ray-tracing results show a much higher shadow fading ( > 10 dB) than mea- surements, due to the larger dynamic range allowed in some ray-tracing experiments. In sub-6 GHz NLoS scenarios, the root mean square (RMS) delay spread is typically modeled in the range of 50-500 ns, the RMS azimuth angle spread of departure (from the access point) at around 10-30 degrees, and the RMS azimuth angle spread of arrival (at the UE) at around 50-80 degrees [7,8]. There are measurements of the delay spread above 6 GHz which indicate somewhat smaller ranges as the frequency increases, and some measure- ments show the millimeter wave omnidirectional channel to be highly directional in nature. A modified Bessel function of the first kind is a function ((x) which is one of the solutions to the modified Bessel differential equation and is closely related to the Bessel function of the first kind Jn(x). The modified Bessel function of the first kind In(z) can be defined by the contour integral In(z) = Sez/2(1+1)/th-1-1 dt where the contour encloses the origin and is traversed in a counterclockwise direction. In terms of Jn(x), New Radio Access Physical Layer Aspects (Part 1) 291 In UMa test environments, the behavior of LoS path loss is similar to free-space path loss and the NLoS path loss behavior appears n

ot following a certain model over a wide range of frequencies. It was observed that the rate at which the path loss increased with frequency was not linear, as the rate is higher in the lower parts of the spectrum, which can be possi- bly explained as due to diffraction effect that is frequency-dependent and a more dominat- ing component in lower frequencies. However, in higher frequencies, reflections and scattering are relatively the predominant components. From preliminary ray-tracing studies, the channel spreads in terms of delay and angle appear to be weakly dependent on the fre- quency and are generally 2-5 times smaller than the values reported in [7]. The cross-polar scattering in the ray-tracing results tends to increase with increasing frequency due to dif- fuse scattering. In InH deployment scenarios, under LoS conditions, multiple reflections from walls, floor, and ceiling could give rise to wave-guiding effect. The measurements conducted in office scenarios suggest that path loss exponent, based on a 1 m free-space reference distance is typically below 2, leading to relatively less path loss than predicted by the free-space path loss formula. The strength of the wave-guiding effect is variable and the path loss exponent appears to increase slightly with increasing frequency, possibly due to the relation between the wavelength and surface roughness. Measurements of the small-scale channel properties such as angular spread (AS) and delay spread have shown similarities between channels over a very wide frequency range, where the results suggest that the main multipath compo- nents are present at all frequencies with some small variations in magnitudes. Recent studies have shown that polarization discrimination ranges between 15 and 25 dB for indoor mmWave channels with greater polarization discrimination at 73 GHz than at 28 GHz [1-3]. Cross-polarized signal components are generated by reflection and diffraction. It is widely known that the fading correlation characteristic between orthogonally polarize

d antennas has a very low correlation coefficient. Polarization diversity techniques and MIMO systems with orthogonally polarized antennas are developed that employ this fading characteristic. Employing the polarization diversity technique is one solution to improving the received power, and the effect of the technique is heavily dependent on the cross-polarization dis- crimination (XPD) ratio characteristic. Moreover, the channel capacity can be improved by appropriately using the cross-polarization components in MIMO systems. Thus the commu- nication quality can be improved by effectively using the information regarding the cross- polarized waves in a wireless system. XPD is an important characteristic, particularly in Diffuse scattering is a form of scattering that arises from deviation of material structure from that of a per- fectly regular lattice, which appears in experimental data as scattering spread over a wide q - range (diffuse). Diffuse scattering is generally difficult to quantify and is closely related to Bragg diffraction, which occurs when scattering amplitudes add constructively. The defect in a crystal lattice results in reduction of the ampli- tude of the Bragg peak. 292 Chapter 3 dual-polarized systems, where cross-talk between polarizations can prevent the system to achieve its quality objectives. Radios can use cross-polarization interference cancellation (XPIC) to isolate polarizations and compensate for any link or propagation induced cou- pling. However, good antenna polarization is important to allow the XPICs maximum flexi- bility to compensate for these dynamic variations. The following equations provide one of the path loss models for UMa test environment for LoS and NLoS scenarios in the frequency range of 0.5-100 GHz. These models were developed based on measurement results conducted by independent academic and industry organizations and published in the literature [7,24,34-36]. LoS path loss for 0.5 GHz <f 100 GHz PLUMa-LoS PL2 = 40 log 10(d3D) + hgNB = 25 m, NLoS path loss for 6

GHz f100 GHz PLUMa-NLoS = 10 m Sd2D 5 5 km PL'UMa-NLoS 13.54 39.08 log 10(d3D) + logio(c) - 0.6(hue - 1.5) 1.5 hue < 22.5 m, hgNB = 25 OSF 6 3.1.1.2 Delay Spread Time-varying fading due to scattering objects or transmitter/receiver motion results in Doppler spread. The time spreading effect of small-scale or microscopic fading is mani- fested in the time domain as multipath delay spread and in the frequency domain as channel coherence bandwidth. Similarly, the time variation of the channel is characterized in the time domain as channel coherence time and in the frequency domain as Doppler spread. In a fading channel, the relationship between maximum excess delay time Tm and symbol time Ts can be viewed in terms of two different degradation effects, that is, frequency-selective fading and frequency non-selective or flat fading. A channel is said to exhibit frequency- selective fading if Tm > Ts. This condition occurs whenever the received multipath compo- nents of a symbol extend beyond the symbol's time duration. Such multipath dispersion of the signal results in inter-symbol interference (ISI) distortion. In the case of frequency- selective fading, mitigating the distortion is possible because many of the multipath compo- nents are separable by the receiver. A channel is said to exhibit frequency non-selective or New Radio Access Physical Layer Aspects (Part 1) 293 flat fading if Tm <Ts. In this case, all multipath components of a symbol arrive at the receiver within the symbol time duration; therefore, the components are not resolvable. In this case, there is no channel-induced ISI distortion, since the signal time spreading does not result in significant overlap among adjacent received symbols. There is still perfor- mance degradation because the irresolvable phasor components can add up destructively to yield a substantial reduction in SNR. Also, signals that are classified as exhibiting flat fad- ing can sometimes experience frequency-selective distortion [33]. Fig. 3.3 illustrates multipath-intensity profil

e A(T) versus delay T where the term delay refers to the excess delay. It represents the signal's propagation delay that exceeds the delay of the first signal component arrival at the receiver. For a typical wireless communication Dual functions Vc V max Vc + V Tm Maximum excess delay Spectral broadening Multipath intensity profile Doppler power spectrum Fourier Fourier transforms transforms Dual functions Spaced-frequency correlation function Spaced-time correlation function Coherence bandwidth Coherence time Figure 3.3 Illustration of the duality principle in time and frequency domains [33]. 294 Chapter 3 channel, the received signal usually consists of several discrete multipath components. The received signals are composed of a continuum of multipath components in some channels such as the tropospheric channel. In order to perform measurements of the multipath- intensity profile, a wideband signal, that is, a unit impulse or Dirac delta function, is used. For a single transmitted impulse, the time Tm between the first and last received component is defined as the maximum excess delay during which the multipath signal power typically falls to some level 10-20 dB below that of the strongest component. Note that for an ideal system with zero excess delay, the function A(T) would consist of an ideal impulse with weight equal to the total average received signal power. In the literature, the Fourier trans- form of A(T) is referred to as spaced-frequency correlation function P(v). The spaced- frequency correlation function P(v) represents the channel's response to a pair of sinusoidal signals separated in frequency by V. The coherence bandwidth Afc is a measure of the fre- quency range over which spectral components have a strong likelihood of amplitude corre- lation. In other words, a signal's spectral components over this range are affected by the channel in a similar manner. Note that fo and Tm are inversely proportional (Afc 1 1/Tm). The maximum excess delay Tm is not the best indicator of how a given wireless

system will perform over a communication channel because different channels with the same value of Tm can exhibit different variations of signal intensity over the delay span. The delay spread is often characterized in terms of the RMS delay spread TRMS in which the average multi- path delay T is calculated as follows: An exact relationship between coherence bandwidth and delay spread does not exist and must be derived from signal analysis of actual signal dispersion measurements in specific channels. If coherence bandwidth is defined as the frequency interval over which the chan- nel's complex-valued frequency transfer function has a correlation of at least 0.9, the coher- ence bandwidth is approximately 2 1/(50TRMS). A common approximation of Afc corresponding to a frequency range over which the channel transfer function has a correla- tion of at least 0.5 is 1/(5TRMS). A channel is said to exhibit frequency-selective effects, if Afc <1/Ts where the inverse symbol rate is approximately equal to the signal bandwidth W. In practice, W may differ from 1/Ts due to filtering or data modulation. Frequency-selective fading effects arise whenever a signal's spectral components are not equally affected by the channel. This occurs whenever Afc < W. Frequency non-selective or flat fading degradation occurs when- ever Afc > W. Hence, all spectral components of the signal will be affected by the channel in a similar manner. Flat fading does not introduce channel-induced ISI distortion, but per- formance degradation can still be expected due to loss in SNR, whenever the signal New Radio Access Physical Layer Aspects (Part 1) 295 experiences fading. In order to avoid channel-induced ISI distortion, the channel is required to exhibit flat-fading by ensuring that Afc > W. Therefore, the channel coherence bandwidth Afc sets an upper limit on the transmission rate that can be used without incorporating an equalizer in the receiver [33]. 3.1.1.3 Doppler Spread Fig. 3.3 shows another function P(t) known as spaced-time correlation f

unction, which is the autocorrelation function of the channel's response to a sinusoid. This function specifies the extent to which there is correlation between the channel's response to a sinusoid sent at time t1 and the response to a similar sinusoid sent at time t2, where At = t2 - t1. The coher- ence time is a measure of the expected time duration over which the channel's response is essentially invariant. To estimate P(t), a sinusoidal signal is transmitted through the chan- nel and the autocorrelation function of the channel output is calculated. The function P(t) and the coherence time T provide information about the rate of fading channel variation. Note that for an ideal time-invariant channel, the channel's response would be highly cor- related for all values of At and P(t) would be a constant function. If one ideally assumes uniformly distributed scattering around a mobile station with linearly polarized antennas, the Doppler power spectrum (i.e., the inverse Fourier transform of spaced-time correlation function A(v)) has a U-shaped distribution as shown in Fig. 3.3. In a time-varying fading channel, the channel response to a pure sinusoidal tone spreads over a finite frequency range Vc - Vmax + Vmax, where Vc and Vmax denote the frequency of the sinusoidal tone and the maximum Doppler spread, respectively. The RMS bandwidth of A(v) is referred to as Doppler spread and is denoted by VRMS that can be estimated as follows: VRMS = The coherence time is typically defined as the time lag for which the signal autocorrelation coefficient reduces to 0.7. The coherence time is inversely proportional to Doppler spread Tc 2 VRMS. A common approximation for the value of coherence time as a function of Doppler spread is T = 0.423/VRMS. It can be observed that the functions on the right side of Fig. 3.3 are dual of the functions on the left side (duality principle). 3.1.1.4 Angular Spread The angle spread refers to the spread in AoA of the multipath components at the receiver antenna array. At the transmitter, on th

e other hand, the angle of spread refers to the spread in the angle of departure (AoD) of the multipath components that leave the transmit 296 Chapter 3 antennas. If the angle spectrum function O(A) denotes the average power as function of AoA, then the RMS angle spread can be estimated as follows: The angle spread causes space-selective fading, which manifests itself as variation of signal amplitude according to the location of antennas. The space-selective fading is characterized by the coherence distance Dc which is the spatial separation for which the autocorrelation coefficient of the spatial fading reduces to 0.7. The coherence distance is inversely propor- tional to the angle spread Dc 1/ORMS. In Fig. 3.3 a duality between multipath-intensity function A(T) and Doppler power spectrum A(v) is shown, which means that the two func- tions exhibit similar behavior across time domain and frequency domain. As the A(T) func- tion identifies expected power of the received signal as a function of delay, A(v) identifies expected power of the received signal as a function of frequency. Similarly, spaced- frequency correlation function (f) and spaced-time correlation function (t) are dual func- tions. It implies that as (f) represents channel correlation in frequency, (t) corresponds to channel correlation function in time in a similar manner [33]. The AS in radians can be expressed based on the circular standard deviation in directional statistics using the following expression [24,25]: where Pnm denotes the power of the mth subpath of the nth path and 4nm is the subpaths angle (either AoA, AoD, elevation angle of arrival [EoA], elevation angle of departure [EoD]). In order to model large signal bandwidths and large antenna arrays, the channel models have been specified with sufficiently high resolution in the delay and angular domains. There are two important aspects related to large antenna arrays. One is the very large size of the antenna array and the other is the large number of antenna elements. These features re

quire high angu- lar resolution in channel modeling, which means more accurate modeling of AoA/AoD, and possibly higher number of multipath components. 3.1.1.5 Blockage The blockage model describes a phenomenon where the stationary or moving objects stand- ing between the transmitter and receiver dramatically change the channel characteristics and in some cases may block the signal, especially in high-frequency bands, since the signal in mmWave does not effectively penetrate or diffract around human bodies and other objects. Shadowing by these objects is an important factor in the link budget calculations and the time variation of the channel, and such dynamic blocking may be important to capture in New Radio Access Physical Layer Aspects (Part 1) 297 evaluations of technologies that include beam-finding and beam-tracking capabilities. The effect of the blockage is considered not only on the total received power, but also on the angle or power of multipath due to different size, location, and direction of the blocker. There are two categories of blockage: (1) dynamic blockage and (2) geometry-induced blockage. Dynamic blockage is caused by the moving objects in the communication envi- ronment. The effect is additional transient loss on the paths that intercept the moving objects. Geometry-induced blockage, on the other hand, is a static property of the environ- ment. It is caused by objects in the map environment that block the signal paths. The propa- gation channels in geometry-induced blockage locations are dominated by diffraction and sometimes by diffuse scattering. The effect is an additional loss beyond the normal path loss and shadow fading. Compared to shadow fading caused by reflections, diffraction- dominated shadow fading may have different statistics (e.g., different mean, variance, and coherence distance) [33]. Radio waves are attenuated by foliage, and this effect increases with frequency. The main propagation phenomena involved are attenuation of the radiation through the foliage, diffraction abov

e/below and sideways around the canopy, and diffuse scattering by the leaves. The vegetation effects are captured implicitly in the path loss equations. A stochastic method for capturing human and vehicular blocking in mmWave frequency regions can be used (among other methods) to model the blockage effect. In this case, the number of blockers must be first determined. For this purpose, multiple 2D angular block- ing regions, in terms of center angle, azimuth, and elevation angular span are generated around the UE. There is one self-blocking region and K = 4 non-self-blocking regions, where K may be changed for certain scenarios such as higher blocker density. Note that the self-blocking component of the model is important in capturing the effects of human body blocking. In the next step, the size and location of each blocker must be generated. For self- blocking, the blocking region in the UE local coordinate system4 is defined in terms of ele- vation and azimuth angles and azimuth and elevation angular span (xsb,Ysb) [4,5]. where the parameters of the above equation are described in Table 3.1. For non-self-blocking k = 1, 2, 3, 4, the blocking region in global coordinate system is defined as follows [4,5]: (0)(0 where d is the distance between the UE and the blocker; other parameters are given in Table 3.1. Global and local coordinate systems are used to locate geometric items in space. By default, a node coordi- nates are defined in the global Cartesian system. 298 Chapter 3 Table 3.1: Blocking region parameters [4,5]. Self-Blocking Region Parameters Portrait (degree) Landscape (degree) Blocking Region Parameters Blocker Index k = 1,2,3, 4 InH scenario Uniform in [0°, Uniform in [15°, Uniform in [5°,15°] UMi_x, Uma_x, RMa_x Uniform in [0°, Uniform in [5°,15°] scenarios The attenuation of each cluster due to self-blocking corresponding to the center angle pair (Osb,L'sb) is 30 dB provided that P'AOA - and - otherwise, the attenuation is 0 dB. Note that ZOA denotes the Zenith angle of arrival. The attenuation of

each cluster due to the non-self-blocking regions k=1,2,3,4 is given as Ldb - 20 log10[1 - provided that and OPOA - Ok <yk; otherwise, the attenuation is 0 dB. The terms FA1|A2|Z||Z2 in the previous equation are defined as follows [4,5]: where A1=4AOA (4k+xx/2), A2 - and Z = OZOA (Ok - yk/2). The center of the blocker is generated based on a uniformly distrib- uted random variable, which is temporally and spatially consistent. The 2D autocorrelation function R(Ax, At) can be described with sufficient accuracy by the exponential function R(Ax, At) where dcorr denotes the spatial correlation dis- tance or the random variable determining the center of the blocker and tcorr is the correlation time given as tcorr = dcorr/v, in which V is the speed of moving blocker [4,5]. 3.1.1.6 Oxygen Absorption The electromagnetic wave may be partially or totally attenuated by an absorbing medium due to atomic and molecular interactions. This gaseous absorption causes additional loss to the radio wave propagation. For frequencies around 60 GHz, additional loss due to oxygen absorption is applied to the cluster responses for different center frequency and bandwidth correspondingly. The additional loss OLn(fc) for cluster n at center frequency fc is given as follows [4]: New Radio Access Physical Layer Aspects (Part 1) 299 Table 3.2: Oxygen attenuation a(fc) as a function of frequency [4,5]. Frequency (GHz) a(fc)(dB/km) where a(fc) denotes the frequency-dependent oxygen absorption loss in dB/km whose sample values at some frequencies are shown in Table 3.2, C is the speed of light in m/s, d3D is the 3D distance in meters between the receive and transmit antennas, Tn is the nth cluster delay in seconds, and TA 0 in the LoS scenarios. For center frequencies not shown in Table 3.2, the frequency-dependent oxygen absorption loss a(fc) is obtained from a linear interpolation of the values corresponding to the two adjacent frequencies [4,5]. For wideband channels, the time-domain channel response of each cluster (all rays within one cluste

r share common oxygen absorption loss) are transformed into frequency-domain channel response and the oxygen absorption loss is applied to the cluster's frequency- domain channel response for frequency fc + Af within the channel bandwidth W. The oxy- gen loss OLn(fc + Af) for cluster n at frequency fc + Af, where W/2 < W/2 is given as follows: where alf + Af) is the oxygen absorption loss in dB/km at frequency fc + Af. The final frequency-domain channel response is obtained by the summation of frequency-domain channel responses of all clusters. Time-domain channel response is obtained by the reverse transform from the obtained frequency-domain channel response [4,5]. Measurements in mmWave frequency bands have shown that ground reflection in mm Wave has significant effect which can produce a strong propagation path that superimposes with the direct LoS path and induces severe fading effects. When ground reflection is considered, the randomly generated shadow fading is largely replaced by deterministic fluctuations in terms of distance. As a result, the standard deviation of shadow fading, when ground reflec- tion is considered, is set to 1 dB. The value of 1 dB was obtained via simulations in order to maintain a similar level of random channel fluctuations without ground reflection [4,5]. The mmWave channels are sparse, that is, they have few entries in the delay angle bins, although experimental verification of this may be limited due to the resolution of rotating horn antennas used for such measurements. However, a lower bound on the channel sparsity can still be established from existing measurements, and in many environments, the percent- age of delay/angle bins with significant energy is rather low but not necessarily lower than at centimeter-wave frequencies. 300 Chapter 3 Molecular oxygen absorption around 60 GHz is particularly high ( ~ 13 dB/km, depend- ing on the altitude) but decreases rapidly away from the oxygen resonance frequency to below 1 dB/km. While these absorption values are considered high f

or macrocell links, cell densification has already reduced the required link distance to a substantially smaller range in urban areas, and the densification process will continue to reduce cell sizes. For a link distance of approximately 200 m, the path loss of a 60 GHz link in heavy rain con- dition is less than 3 dB. Fig. 3.4 illustrates the impact of rainfall and oxygen absorption throughout the nmWave transmission. A more serious issue than free-space loss for nmWave signals is their limited penetration through materials and limited diffraction. In the urban environment, coverage for large cells could be particularly challenging; how- ever, cell densification can be used to achieve the ambitious 5G capacity goals. Atmospheric effects such as oxygen and water vapor absorption as well as fog and precipi- tation can scale exponentially with the link distance. Limiting to distances below 1 km, attenuation caused by atmospheric gases can be neglected up to 50 GHz, as shown in Fig. 3.4. However, above 50 GHz, it becomes important to consider the oxygen absorption peak of approximately 13 dB/km at 60 GHz and the water vapor resonance peak at 183 GHz of approximately 29 dB/km for relative humidity of 44% under standard Gaseous H2O and O2 Free space Frequency (GHz) Figure 3.4 Comparison of gaseous H2O + O2, rain, and free-space attenuation (propagation distance 1 km, rain rate of 95 mm/h, and dry air pressure is 1013 hPa and the water vapor density is 7.5 g/m³. New Radio Access Physical Layer Aspects (Part 1) 301 conditions. Note that neither fog nor rain is relevant for frequencies below 6 GHz. At fre- quencies above 80 GHz, dense fog related to a visibility of less than 70 m has a noticeable impact dB/km) and becomes severe above 200 GHz ( > 10 dB/km). Drizzle and steady rain are not a substantial issue for distances up to 1 km (3.0-4.4 dB/km above 70 GHz). However, as shown in Fig. 3.4, heavy rain attenuation increases up to 10-15 dB/km, and for downpours, up to 40 dB/km can be experienced. In summary, atmospheric

effects, espe- cially under bad weather conditions, are relevant for mmWave links over distances greater than 100 m and a crucial issue for longer distances of 1 km and farther [32]. 3.1.1.7 LoS Path Loss Probability It is shown in the literature that a height-dependent path loss for an indoor UE associated with a LoS condition can be modeled considering the dimensions of the building and the location of the UE inside the building. 3GPP has modeled the LoS path loss by using the 3D distance between the gNB and the UE along with the coefficients given by the ITU-R LoS path loss equations for 3D-UMa and 3D-UMi [4,24]. This provides a reasonable approximation to the more accurate models and can be determined without explicitly model- ing the building dimensions. The ITU-R LoS path loss model assumes a two-ray model resulting in a path loss equation transitioning from a 22 dB/decade slope to a steeper slope at a break point depending on the environment height, which represents the height of a dominant reflection from the ground (or a moving platform) that can add constructively or destructively to the direct ray received at a UE located at the street level. In the 3D-UMa scenario, it is likely that such a dominant reflection path may come from the street level for indoor UEs associated with a type-1 LoS condition. Therefore, the environmental height is fixed at 1 m for a UE associated with a type-1 LoS condition. In the case of a UE associated with a type-2 LoS condition a dominant path can be likely created by reflection from the rooftop of a neighboring building. Note that a rooftop is at least 12 m in height; in this case the environmental height is randomly determined from a discrete uniform distribution between the UE height in meter and 12 [4,24]. For NLoS path loss modeling, which is the primary radio propagation mechanism in a 3D- UMa scenario, the dominant propagation paths experience multiple diffractions over roof- tops followed by diffraction at the edge of the building. The path loss attenuation increase

the composite power angular spectrum in zenith statistically has a Laplacian distribution and its conditional distribution given a certain link distance and UE height can also be approximated by Laplacian distribution. To incorporate these observations, zenith angle of departure (ZoD) and zenith angle of arrival (ZoA) are modeled by inverse Laplacian func- tions. It is also observed that the zenith angle spread of departure (ZSD) decreases signifi- cantly as the UE moves further away from the gNB. An intuitive explanation is that the angle subtended by a fixed local ring of scatterers at the UE to the gNB decreases as the UE moves away from the gNB. The ZSD is also observed to slightly change as an indoor UE moves up to higher floors [34-36]. We use the concept of the LoS probability to distinguish between the LoS and NLoS links. A link of length d is LoS with probability PLos(d). The LoS probability is a non-increasing function of the link length. The LoS probability is obtained based on the certain building models, that is, either we use stochastic models from random shape theory or we use site- specific maps from geographical information system database. In this section, the LoS con- dition is determined based on a map-based approach, that is, by considering the transmitter and receiver positions and whether any buildings or walls are blocking the direct path between them. The impact of in-between objects not represented in the map is modeled sep- arately through shadowing or blocking path loss components. It is noteworthy that this LoS definition is frequency independent, due to the fact that only buildings and walls are consid- ered in the definition. 3GPP and ITU-R define the UMa LoS probability as follows: where d2D is the 2D distance in meters and d1 and d2 can be optimized to fit a set of mea- surement data in the test environments/scenarios under consideration. For UMi test environ- ments, it was observed that the above LoS probability is sufficient for frequencies above 6 GHz. The fitted d1/d2 model pr

ovides more consistency with measured data and the error between the measured data and the 3GPP LoS probability model over all distances are small. Note that the 3GPP UMi LoS probability model is not a function of UE height unlike the UMa LoS probability model. New Radio Access Physical Layer Aspects (Part 1) 303 d2D-out d2D-in Outdoor scenario Indoor scenario Figure 3.5 Definition of 2D and 3D distances in outdoor/indoor environments [4]. RMa (021) UMa (O2I) UMi (O2I) InH-Office (NLOS) RMa (NLOS) UMa (NLOS) UMi (NLOS) InH-Office (NLOS-opt) RMa (LOS) UMa (NLOS-opt) UMi (NLOS-opt) InH-Office (LOS) UMa (LOS) UMi (LOS) InH-Shopping mall (LOS) 3D distance (m) Figure 3.6 Path loss with (solid line) and without (dashed line) shadowing for various scenarios [37]. In the path loss models we often refer to 2D and 3D distances. Those distances are defined as follows (see Fig. 3.5): Fig. 3.6 shows the path loss in dB for the 3D distance from the smallest value supported in each scenario to 103 m for the outdoor scenarios and 102 m for the indoor cases. In this fig- ure, O2I denotes outdoor-to-indoor loss in various test environments. 3.1.2 Two- and Three-Dimensional Channel Models The 5G cellular systems are expected to operate over a wide range of frequencies from 450 MHz to 100 GHz. For the development and standardization of the new 5G systems operating in frequency bands above 6 GHz, there was a serious need to accurately model radio signal propagation in these frequency bands, which could not be fully characterized by the existing channel models, because the previous generations of channel models were designed and evaluated for sub-6 GHz frequencies. The development of 3GPP 3D channel 304 Chapter 3 model was a step toward modeling 2D antenna arrays that are used in 5G network deploy- ments. The measurements indicate that the smaller wavelengths increase the sensitivity of the propagation models to the scale of the environment effects and show some frequency depen- dence of the path loss as well as increased occurrence o

f blockage phenomenon. Furthermore, the penetration loss is highly dependent on the material and tends to increase with frequency. The shadow fading and angular spread parameters are larger and the boundary between LoS and NLoS depends not only on antenna heights but also on the local environment. The small- scale characteristics of the channel such as delay-spread and angular-spread and the multipath richness is somewhat similar over the, which was a good reason for extending the existing 3GPP models to the wider frequency range [24,25]. The goal of channel modeling is to provide accurate mathematical representations of radio propagation to be used in link-level and system-level simulations corresponding to a specific deployment scenario. Since the radio channel can be assumed as linear, it can be described by its impulse response. Once the impulse response is known, one can determine the response of the radio channel to any input signal. The impulse response is usually represented as a power density function of excess delay, measured relative to the first detectable signal. This function is often referred to as a power delay profile. The channel impulse response varies with the position of the receiver and may also vary with time. Therefore, it is usually measured and reported as an average of profiles measured over one wavelength to reduce noise effects, or over several wavelengths to determine a spatial average. It is important to clarify which aver- age is meant, and how the averaging was performed. The propagation effects of a wireless channel can be modeled with a large-scale propagation model combined with a small-scale fading model, where the former models long-term slow- fading characteristics of the wireless channel, such as path loss and shadowing, while the small-scale fading model provides rapid fluctuation behavior of the wireless channel due to multipath and Doppler spread. For a wireless channel with multiple antennas, static beam- forming gain such as the sectorization beam pattern also contribu

tes to the long-term propa- gation characteristic of the wireless channel and can be modeled as part of the large-scale propagation model. As for the small-scale fading model of a MIMO channel, the correlation of signals between antenna elements also needs to be considered and can be modeled by a spatial channel model that was used to evaluate performance of the previous generations of cellular standards. Although electromagnetic beam patterns generated by base station antenna arrays are 3D in nature, they were usually modeled as linear horizontal arrays, and elevation angles of signal paths have been ignored for simplicity. The 3D spatial channel models non-zero elevation angles associated with signal paths as well as azimuth angles SO that the small-scale fading effect on each antenna element of the 2D antenna grid and the correlation between any two pair of antenna elements on the 2D antenna grid can be modeled. New Radio Access Physical Layer Aspects (Part 1) 305 While the increase in average mutual information (capacity) of a MIMO channel with the number of antennas is well understood, it appears that the variance of the mutual informa- tion can grow very slowly or even shrink as the number of antennas increases. This phe- nomenon is referred to as channel hardening in the literature, which has certain implications for control and data transmission [31]. For a three-sector macrocell, the horizontal and vertical antenna patterns are commonly modeled with a 3 dB beam-width of 70 and 10 degrees, respectively, with an antenna gain of 17 dBi. The vertical antenna pattern is also a function of electrical and mechanical antenna downtilt, where the electrical downtilt is a result of vertical analog beamforming, generated by applying common and static phase shifts to each vertical array of 2D antenna elements. For instance, the electrical downtilt is set to 15 and 6 degrees for a macrocell deployment scenario with inter-site distance of 500 m and 1732 m, respectively, in 3GPP evaluation methodology [4]. In order to m

eet the technical requirements of IMT-2020, new features are captured in 5G channel models such as support of frequencies up to 100 GHz and large bandwidth, 3D modeling, support of large antenna arrays, blockage modeling, and spatial consistency [4,5]. The 3D modeling describes the channel propagation in azimuth and elevation directions between the transmitting and the receiving antennas. It is more complete and accurate rela- tive to the 2D modeling which only considers the propagation characteristics in the azimuth direction. Multi-antenna techniques capable of exploiting the elevation dimension have been developed for LTE since Rel-13 and are considered very important in 5G, which include modeling the elevation angles of departure and arrival, and their correlation with other para- meters [7]. For a base station (access node) equipped with columns of active antenna arrays, vertical analog beamforming can be applied prior to the power amplifier (PA); thus the electrical downtilt can be dynamically adjusted over time. This feature allows the base sta- tion to dynamically adjust its cell coverage depending on the user distribution in the cell, for example, analog beamforming can be directed toward UEs congregated at the same location. Nonetheless, such beamforming is still fundamentally cell specific (i.e., spatial separation of individual UE is not possible) and operate on a long timescale, although time- varying base station antennas are modeled having one or multiple antenna panels, where an antenna panel has one or multiple antenna elements placed vertically, horizontally or in a 2D array within each panel. As a result, 3D channel modeling is required for performance evaluation of full-dimension MIMO. The full-dimension MIMO involves precoding/beam- forming exploiting both horizontal and vertical degrees of freedom on a small timescale in a frequency-selective manner. We start our study with the 3D antenna models in mmWave bands where implementation of large antenna arrays is feasible in the gNB or the UE. Le

t's assume that the gNB and/or 306 Chapter 3 Cross-polarized elements Co-polarized elements Figure 3.7 3D base station antenna model [4]. the UE each has a 2D planar antenna subarray comprising Ms X Ns antenna elements, where Ns denotes the number of columns and Ms is the number of antenna elements with the same polarization in each column (see Fig. 3.7). The antenna elements are uniformly spaced with a center-to-center spacing of dh and dv in the horizontal and vertical directions, respectively. The Ms X Ns elements may either be single polarized or dual polarized. A uniform rectangu- lar array is formed comprising MXN antenna panels where M is the number of panels in a column and N is the number of panels in a row. Antenna panels are uniformly spaced with a center-to-center spacing of dH and dv in the horizontal and vertical directions, respec- tively. The 3GPP 3D channel model allows modeling 2D planar antenna arrays. The antenna elements can either be linearly polarized or cross polarized, as shown in Fig. 3.7. In this regard, the model represents a compromise between practicality and precision as it does not include the mutual coupling effect as well as different propagation effects of horizon- tally and vertically polarized waves [7]. For each antenna element, the general form of antenna element horizontal radiation pattern can be specified as AE.H(O) = - min ( where - 180° O < 180°, denotes the horizontal 3 dB beam-width, and the parameter am is the maximum side-lobe level attenuation. The general form of antenna element vertical radiation pattern is specified as Ae,v( = - min (12 where 180° O V 180°, A3dB denotes the verti- cal 3 dB beam-width, and Otilt is the tilt angle. It must be noted that 0=0 points to the zenith and O =90° points to the horizon. The combined vertical and horizontal antenna New Radio Access Physical Layer Aspects (Part 1) 307 element pattern is then given as A(A, - - where A(A, 0) is the relative antenna gain (dB) of an antenna element in the direction (0,0) [4,7,24]. The above conc

epts can also be applied to the UE antenna arrays. In this case MXN antenna panels may have different orientations. Let denote the orientation angles of the panel (mg,ng) < 0 ng <Ng, where the orientation of the first panel (520,0,00,0) is defined as the UE orientation, Smagng is the array bearing angle, and Omg,ng is the array downtilt angle. The antenna bearing is defined as the angle between the main antenna lobe center and a line directed toward east. The bearing angle increases in a clockwise direction. The parameters of the base station antenna array pattern for Dense- urban-eMBB (macro-TRP), Rural-eMBB, Urban-macro-mMTC, and Urban-macro-URLLO test environments (i.e., deployment scenarios studied in 3GPP and ITU-R5) are defined as 03dB = = 65°, Otilt==00, = am = 30 dB. The parameters of the UE antenna pattern for frequencies above 30 GHz are given as 03dB = =90°, Otilt==00, = and am = 25 dB. The IMT-2020 3D channel model developed in ITU-R is a geometry-based stochastic chan- nel model. It does not explicitly specify the location of the scatterers, rather the directions of the rays. Geometry-based modeling of the radio channels enables separation of propaga- tion parameters and antennas. The channel parameters for individual snapshots are deter- mined stochastically based on statistical distributions obtained from channel measurements. Channel realizations are generated through the application of the geometrical principle by summing contributions of rays with specific small-scale parameters such as delay, power, azimuth angles of arrival and departure, and elevation angles of arrival and departure. The results are further combined with transmit/receive antenna correlations and temporal fading with geometry-dependent Doppler spectrum effects [4]. A single-link channel model is shown in Fig. 3.8 for the downlink direction. Each circle with several dots represents scat- tering region creating one cluster, each cluster is constituted by M rays, where a total of N clusters are assumed. We assume that there are N

tx antennas at the transmitter and Nrx antennas at the receiver. The small-scale parameters such as delay Tn,m, azimuth AoA Test environment reflects geographic environment and usage scenario which is used for the evaluation pro- cess; however, it has a direct relevance to the deployment scenario. The test environments defined in 3GPP and ITU-R are as follows: Indoor hotspot-eMBB is an indoor isolated environment at offices and/or in shopping malls based on stationary and pedestrian users with very high user density. Dense-urban-eMBB is an urban environment with high user density and traffic loads focusing on pedes- trian and vehicular users. Rural-eMBB is a rural environment with larger and continuous wide area coverage, supporting pedes- trian, vehicular, and high speed vehicular users. Urban macro-mMTC is an urban macro-environment targeting continuous coverage focusing on a high number of connected machine type devices. Urban macro-URLLC is an urban macro-environment targeting ultra-reliable and low-latency communications. 308 Chapter 3 Paths Elevation Cluster rx,n.m Azimuth tx,n,m Cluster n rx,n,m Ray m Fix,nov, ) Transmitter antenna pattern Frx,nm,9 ( ) Receiver antenna pattern ZoD offset LoS ZoD ZoD mean Diffraction Penetration Outdoor LoS UEs Outdoor/indoor NLoS UEs Refraction Reflection Figure 3.8 Illustration of 3D MIMO channel model and its parameters [4]. New Radio Access Physical Layer Aspects (Part 1) 309 Orx,n,m' elevation AoA Orx,n,m, azimuth AoD Qtx,n,m and elevation AoD Otx,n,m are assumed to be different for each ray. In the primary module of 3GPP 3D channel model, the number of clusters are fixed and frequency independent. The typical number of clusters reported in the literature is often small, random, and can be modeled as a Poisson distribution. By choosing an appropriate mean value of the Poisson distribution, the events with a larger number of clusters with a low probability may also be produced. The 3GPP 3D channel model is a geometric stochastic model, describing the scattering environ-

ment between base station sector and the UE in both azimuth and elevation dimensions. The scatterers are represented by statistical parameters without having a real physical location. It specifies three propagation conditions, LoS, non-LoS, and outdoor-to-indoor. In each scenario it defines different parameters for mean propagation path loss, macroscopic fading, and micro- scopic fading. The probability of being in LoS is determined separately for indoor and outdoor UEs depending on the height of the UE as well as the break point distance. The break point dis- tance characterizes the gap between transmitter and receiver at which the Fresnel zone is barely broken for the first time. For an indoor UE, LoS refers to the signal propagation outside of the building in which the UE is located. For each UE location, large-scale parameters are generated according to its geographic position as well as the propagation conditions at this location. The large-scale parameters incorporate shadow fading, the Rician K-factor (only in the LoS case), delay spread, azimuth angle spread of departure and arrival, as well as azimuth angle-spread of departure and arrival. The time-variant impulse response matrix of Nrx X Ntx MIMO channel is given by the following expression [4]: where t, T, N, and n denote time, delay, number of clusters, and cluster index, respectively. The above channel impulse response is composed of the antenna array impulse response matrices Fix and Frx for the transmitter and the receiver sides, respectively, as well as the dual-polarized propagation channel response matrix. The channel from the transmitter antenna element Nix to the receiver antenna element Nrx for the cluster n is given by the fol- lowing expression [4]: exp (janm) exp (jpn) 0n,m,AoD) exp(p 310 Chapter 3 where Pn is the power of nth path and the other parameters are defined as follows [4]: In and °C n,m denote random initial phase for each ray m of each cluster n and for four different polarization combinations. Frx,Nrx,9 and Frx,Nnx,O represen

t receive antenna element Nrx field patterns in direction of the spherical basis vectors and . Ftx,mx,0 and Ftx,Nx,0 denote transmit antenna element Ntx field patterns in direction of the spherical basis vectors and . r rx,n,m and r tx,n,m denote spherical unit vector with azimuth arrival angle eleva- tion arrival angle On,m,ZoA and azimuth departure angle Qn,m,AoD, and elevation departure angle On,m,ZoD. drx,nrx and dix,nx represent location vector of receive antenna element Nrx and transmit antenna element Ntx. Kn,m is the cross-polarization power ratio in linear scale. to denotes the wavelength of the carrier frequency. Vn,m represents the Doppler frequency component of ray n, m. If the radio channel is dynamically modeled, the above small-scale parameters would become time variant. The channel impulse response describes the channel from a transmit antenna element to a receive antenna element. The spherical unit vectors are defined as fol- lows [24]: rx,n,m tx,n,m The Doppler frequency component Vn,m depends on the arrival angles (AoA, ZoA), and the UE velocity vector V with speed V, travel azimuth angle Ov, elevation angle O and is given by The parameters drx,Nrx and dix,nx are the location vectors of receive and transmit antenna elements, respectively. Considering a base station sector with coordinates (Sx,Sy,Sz) and a planar antenna array, the location vector per antenna element is defined as dix,nx = (Sx Sy sz) + ( 0 (k-1)dH (1-1)dy) where = 1,2, , N l=1,2,...,M = (see Fig. 3.7). The parameters N and dH denote the number of antenna ele- ments and the inter-element spacing in the horizontal direction, respectively, while M and dv represent the number of antenna elements and the inter-element spacing in the vertical direction, respectively. The Doppler frequency component of the UE moving at velocity V is represented by parameter Vn,m. New Radio Access Physical Layer Aspects (Part 1) 311 When antenna arrays are deployed at the transmitter and receiver, the impulse response of such arrangement results in a ve

ctor channel. An example of this configuration is given below for the case of a 2D antenna array [4]: exp (jpm exp (jpm) exp(j P)p(2) where atx (On,m,ZoD, On,m,AOD) and arx (On,m,ZoA, On,m,AoA) are the transmit and receive antenna array impulse response vectors, respectively, corresponding to ray me {1, ,M} in cluster ne{1,..., N} given by the following expression: = exp An,m ---- = exp An,m where to is the wavelength of carrier frequency fo; ,, and r rx(On,m,ZoA, are the corresponding angular 3D spherical unit vectors of the transit and receive, respectively; Wix and Wrx denote the location matrices of the transmit and receive antenna elements in 3D Cartesian coordinates. The location matrices in the vectored impulse response are provided for an antenna configuration that is a uniform rectangular array consisting of cross-polarized antenna elements shown in Fig. 3.4. The Doppler shift generally depends on the time-variance of the channel as it is defined as the derivative of the channel phase over time. It can result from transmitter, receiver, or scatterers movement. The general form of the exponential Doppler component is given as follows [24]: where ,,(t) denotes the normalized vector that points to the direction of the incoming wave as seen from the receiver at time t. The velocity vector of the receiver at time t is 312 Chapter 3 denoted by v(t) while to denotes a reference point in time that defines the initial phase ==0. The above expression is only valid for time-invariant Doppler shift, satisfying Spatial correlation is often said to degrade the performance of multi-antenna systems and imposes a limit on the number of antennas that can be effectively fit in a small mobile device. This seems intuitive as the spatial correlation decreases the number of independent channels that can be created by precoding. When modeling spatial correlation, it is useful to employ the Kronecker model, where the correlation between transmit and receive antennas are assumed independent and separable. This model is reasonable

when the main scattering appears close to the antenna arrays and has been validated by outdoor and indoor measure- ments. Assuming Rayleigh fading, the Kronecker model means that the channel matrix can be represented as = (R1/2) where the elements of Hchannel are independent and identically distributed as circular symmetric complex Gaussian random variables with zero- mean and unit variance. The important part of the model is that Hchannel is pre-multiplied by the receive-side spatial correlation matrix RRX and post-multiplied by transmit-side spatial cor- relation matrix RTX. Equivalently, the channel matrix can be expressed as H ~ C(0,RTX x RRX) where x denotes the Kronecker product [9]. Using the Kronecker model, the spatial correlation depends directly on the eigenvalue distri- butions of the correlation matrices RRX and RTX. Each eigenvector represents a spatial direc- tion of the channel and its corresponding eigenvalue describes the average channel/signal gain in that direction. For the transmit-side, the correlation matrix RTX describes the average gain in a spatial transmit direction, while receive-side correlation matrix RRX describes a spatial receive direction. High spatial correlation is represented by large eigenvalue spread in RTX or RRX, implying that some spatial directions are statistically stronger than the others. On the other hand, low spatial correlation is represented by small eigenvalue spread in RTX or RRX, which implies that almost the same signal gain can be expected from all spa- tial directions [9]. Let's now focus on a specific case of downlink transmission from a base station to a mobile station. Denoting the nth snapshot of the spatial correlation matrices at the gNB and the UE by RgNB,n and RUE,n, the per-tap spatial correlation is determined as the Kronecker product of the gNB and UE's antenna correlation matrices as Rn = RgNB,n X RUE,n. We denote the number of receive antennas by NRX and the number of transmit antennas by NTX. If Given a m X n matrix A and a pXq matrix B, the K

ronecker product C = A x B, also called matrix direct product, is an mp X nq matrix with elements defined by cab = ajkl where a = p(i - 1) + k and B = q(j - 1) + 1. The matrix direct product provides the matrix of the linear transformation induced by the vector space tensor product of the original vector spaces. More precisely, suppose that operators S:V1 -> W1 and T:V2 W2 are given by S(x) = Ax and T(y) = By, then S T:V1 x V2 W1 x W2 is determined by SQT(xy) = =(Ax)@(By)=(A@B)(xy). = New Radio Access Physical Layer Aspects (Part 1) 313 Table 3.3: gNB/UE correlation matrix [9]. Number of Antennas Entity RgNB 1 RUE = 1 Rue=(1-1) Table 3.4: Values of a and B for different antenna correlations [9]. Low Correlation Medium Correlation High Correlation cross-polarized antennas are present at the receiver, it is assumed that NRx/2 receive anten- nas have the same polarization, while the remaining NRx/2 receive antennas have orthogo- nal polarization. Likewise, if cross-polarized antennas are present at the transmitter, it is assumed that NTx/2 transmit antennas have the same polarization, while the remaining NTx/2 transmit antennas have orthogonal polarization. It is further assumed that the antenna arrays are composed of pairs of co-located antennas with orthogonal polarization. Under these assumptions, the per-tap channel correlation is determined as Rn = RgNB,n X T X RUE,n where RUE,n is an NRX X NRX matrix, if all receive antennas have the same polarization, or an NRx/2 X NRx/2 matrix, if the receive antennas are cross- polarized. Likewise, RgNB,n is an NTX X NTX matrix, if all transmit antennas have the same polarization, or a NTX/2XXTX/2 matrix, if the transmit antennas are cross-polarized. Matrix T is a cross-polarization matrix based on the cross-polarization defined in the cluster-delay-line models. Matrix T is a 2X2 matrix, if cross-polarized antennas are used at the transmitter or at the receiver. It is a 4 X 4 matrix if cross-polarized antennas are used at both the transmitter and the receiver [9]. Table 3.3

defines the correlation matrices for the gNB and UE in NR with different number of transmit/receive antennas at each entity [9]. For the scenarios with more antennas at either gNB, UE, or both, the channel spatial corre- lation matrix can be expressed as the Kronecker product of R&NB and RUE according to Rspatial = RUE x R&NB. The parameters a and B for different antenna correlation types are given in Table 3.4. The 3D channel model parameters and model generation procedure are summarized in Fig. 3.9. Chapter 3 Measurements Material characteristics Set environment, network layout, and antenna array parameters (v, fc, azimuth o, AoD, AoA, zenith O, ZoD, ZoA) Extend the number Assign propagation condition (NLoS, LoS) of antennas Calculate path loss (dB) Generate large-scale correlation parameters (delay, angular Blocking spread, ricean K, shadow fading) Generate delays (T) Generate cluster powers (P) Ray tracing Generate arrival and departure angles (AoA, ZoA, AoD, ZoD) Coupling of the rays randomly Generate cross-polarization power ratios XPRs (kn,m) Doppler Draw random initial phases Generate channel coefficients (Hnrx,ntx,n(t)) Apply path loss and shadowing Polarization Figure 3.9 Summary of 3D channel model parameters and model generation procedure [7]. 3.2 Waveforms There have been considerable discussions on whether a new type of transmission wave- form besides the incumbent cyclic prefix (CP)-OFDM should be adopted for NR [10,11]. Several alternative OFDM-based waveforms, including filter bank multicarrier and gener- alized frequency division multiplexing, were studied. Many of them claimed advantages in terms of increased bandwidth efficiency, relaxed synchronization requirements, reduced inter-user interference, reduced out-of-band (OOB) emissions, and SO on, but at the same time created challenges in terms of increased transceiver complexity, difficulty in MIMO integration, and significant specification impacts. This section describes some of the prominent 5G waveform candidates and their characteristics

as well as the reasons that the status-quo OFDM waveform continued to be supported in the new radio. Furthermore, the waveform, numerology, and frame structure should be chosen to enable efficient time/frequency utilization for frequency division duplex (FDD) and time division duplex (TDD) deployments, respectively. Table 3.5 summarizes the design requirements concerning the choice of the waveform, which were used to examine the waveform candidates for the NR. Gabor's theory of communication suggests that ideally a multicarrier system such as OFDM must satisfy the following requirements [45]: The subcarriers are mutually orthogonal in time and frequency to make the receiver as simple as possible and to maintain the inter-carrier interference as low as possible. New Radio Access Physical Layer Aspects (Part 1) 315 Table 3.5: Summary of design targets for the waveform [45,47]. Design Criteria Remarks Higher spectral efficiency and High spectral efficiency for high data rates and efficient use of the scalability available spectrum Ability to efficiently support MIMO and multipath robustness Low latency Lower in-band and out-of- Reduce interference among users within allocated band and reduce band emissions interference among neighbor operators Enables asynchronous multiple Support a higher number of small-cell data burst devices with minimal access scheduling overhead through asynchronous operations and enables lower power operation Lower power consumption Low peak-to-average power ratio allowing efficient power amplifier design Lower implementation Reasonable transmitter and receiver complexity and additional complexity complexity must be justified by significant performance improvements Coexistence with legacy and Simplify LTE coexistence mobility Support Robust against Doppler shift to allow high mobility The transmission waveform is well localized in time and frequency. This provides immunity to ISI from multipath propagation or delay spread and to ICI from Doppler spread. A good time localization is required t

o enable low latency. Maximal spectral efficiency, that is, P = (8T8F)-1 with P denoting the spectral effi- ciency in data symbols per second per Hertz. However, it is shown in the literature [38,46], that it is not possible to satisfy these three requirements at the same time and certain tradeoffs are necessary. This con- clusion has an impact on the waveform selection in wireless communication systems. 3.2.1 OFDM Basics and Transmission Characteristics OFDM was selected in LTE/LTE-A due to its efficiency and simplicity using baseband modulation and demodulation stages based on FFT. Mathematically, the nth OFDM symbol can be described as where k denotes the subcarrier index and n denotes symbol index, NFFT is the total number of subcarriers, rect(.) is a rectangular pulse with the symbol period of , and Sk,n is the data symbol [e.g., quadrature amplitude modulation (QAM) symbol] of the kth subcarrier at the nth time instant. A CP is appended to the beginning of each symbol. The inserted CP serves as a guard time between symbols which protects against ISI. In addition, it preserves the orthogonality between subcarriers after passing through a channel provided that the CP 316 Chapter 3 duration is longer than the channel RMS delay spread. The CP acts as a buffer region where delayed information from the previous symbols can be stored. The receiver must exclude samples from the CP which might be corrupted by the previous symbol when choosing the samples for an OFDM symbol. When demodulating the received symbol, the receiver can choose Tu/Ts samples from a region which is not affected by the previous symbol. In a conventional serial data transmission system the information bearing symbols are trans- mitted sequentially, with the frequency spectrum of each symbol occupying the entire avail- able bandwidth. An unfiltered QAM signal spectrum can be described in the form of sin(7fTu)/Tff" with zero-crossing points at integer multiples of 1/Tu, where Tu is the QAM symbol period. The concept of OFDM is to transmit the dat

a bits in parallel QAM- modulated subcarriers using frequency division multiplexing. The carrier spacing is care- fully selected SO that each subcarrier is located on other subcarriers' zero-crossing points in the frequency domain. Although there are spectral overlaps among subcarriers, they do not interfere with each other, if they are sampled at the subcarrier frequencies. In other words, they maintain spectral orthogonality. The OFDM signal in frequency domain is generated through aggregation of NFFT parallel QAM-modulated subcarriers where adjacent subcar- riers are separated by subcarrier spacing 1/Tu. Since an OFDM signal consists of many par- allel QAM subcarriers, the mathematical expression of the signal in time domain can be expressed as follows: (NFFT-1)/2 Siei2nk(t-tg)/Tw k=-(NFFT-1)/2 where x(t) denotes the OFDM signal in time domain, Sk is the complex-valued data that is QAM-modulated and transmitted over subcarrier k, NFFT is the number of subcarriers in frequency domain, Wc is the RF carrier frequency, and Tg is the guard interval or the CP length. For a large number of subcarriers, direct generation and demodulation of the OFDM signal would require arrays of coherent sinusoidal generators which can become excessively complex and expensive. However, one can notice that the OFDM signal is actually the real part of the inverse discrete Fourier transform (IDFT) of the original complex-valued data symbols {sk|k = - (NFFT 1)/2, (NFFT 1)/2}. It can be observed that there are N < NFFT subcarriers each carrying the corresponding data Ak. The inverse of the subcarrier spacing Af = 1/Tu is defined as the OFDM useful symbol dura- tion , which is NFFT times longer than that of the original input data symbol duration. 3.2.1.1 Cyclic Prefix The inclusion of CP in OFDM makes it robust to timing synchronization errors. Robustness to synchronization errors is relevant when synchronization is hard to achieve such as over New Radio Access Physical Layer Aspects (Part 1) 317 the sidelink. It can also be relevant if a

synchronous transmissions are allowed in the uplink. The inclusion of he cyclic prefix adds redundancy to the transmission since the same con- tent is transmitted twice as the CP is a copy of the tail of a symbol placed at its beginning. This overhead can be expressed as a function of symbol duration and duration of the CP as OH = TCP/(TCP OFDM is a flexible waveform that can support diverse services in a wide range of frequencies when properly selecting subcarrier spacing and CP. Further dis- cussion on OFDM numerology design that fulfills a wide range of requirements is given in the next section. Since IDFT is used in the OFDM modulator, the original data are defined in the frequency domain, while the OFDM signal s(t) is defined in the time domain. The IDFT can be implemented via a computationally efficient FFT algorithm. The orthogonality of subcar- riers in OFDM can be maintained and individual subcarriers can be completely separated and demodulated by an FFT at the receiver when there is no ISI introduced by communi- cation channel. In practice, linear distortions such as multipath delay cause ISI between OFDM symbols, resulting in loss of orthogonality and an effect that is similar to cochan- nel interference. However, when delay spread is small, that is, within a fraction of the OFDM useful symbol length, the impact of ISI is negligible, although it depends on the order of modulation implemented by the subcarriers (see Fig. 3.10). A simple solution to mitigate multipath delay is to increase the OFDM effective symbol duration such that it is much larger than the delay spread; however, when the delay spread is large, it requires a large number of subcarriers and a large FFT size. Meanwhile, the system might become sensitive to Doppler shift and carrier frequency offset. An alternative approach to mitigate multipath distortion is to generate a cyclically extended guard interval, where each OFDM symbol is prefixed with a periodic extension of the signal itself, as shown in Fig. 3.10 where the tail of the symbo

l is copied to the beginning of the symbol. The OFDM symbol duration then is defined as Ts = Tu + T, where Tg is the guard interval or CP. When the guard interval is longer than the channel impulse response or the multipath delay, the ISI can be effectively eliminated. The ratio of the guard interval to useful OFDM symbol duration depends on the deployment scenario and the frequency band. Since the insertion of the guard intervals will reduce the system throughput, Tg is usually selected less than Tu/4. The CP should absorb most of the signal energy dispersed by the multipath channel. The entire the ISI energy is contained within the CP, if its length is greater than that of the channel RMS delay spread (Tg >TRMS). In general it is sufficient to have most of the delay spread energy absorbed by the guard interval, considering the inherent robustness of large OFDM symbols to time dispersion. Fig. 3.11 illustrate the OFDM modulation and demodulation process in the transmitter and the receiver, respectively. In practice, a win- dowing or filtering scheme is utilized in the OFDM transmitter side to reduce the OOB emissions of the OFDM signal (see Fig. 3.10). OFDM symbol 1 Corrupted due to ISI Received waveform Delayed OFDM symbol 1 IFFT output OFDM symbol 2 Weighting Weighting function Weighted received waveform function Interference region Overlapped Overlapped extension extension Overlap and add OFDM symbol 1 Delayed OFDM symbol 1 Transmitted waveform FFT input OFDM symbol 2 Interference region Cyclic prefix concept Weighted overlap and add at transmitter side Weighted overlap and add at receiver side Figure 3.10 Illustration of the effect of cyclic prefix for eliminating ISI (h(t) is the hypothetical channel impulse response) and practical implementation of CP insertion and removal 47]. New Radio Access Physical Layer Aspects (Part 1) 319 S-(N-1)/2 S-(N-1)/2+1 S-(N-1)/2+2 cos(w) Channel Serial to Parallel to coding and parallel serial Cyclic prefix Windowing conversion insertion and filtering modulation converter c

onversion and filtering S(N-1)/2-2 S(N-1)/2-1 Modulated S(N-1)/2 (m+ 1)T, Input bits symbols Xm 'm (N-1 Transmitter One useful OFDM symbol Frequency domain Time domain Discard Serial to Frequency Parallel to CP removal parallel domain serial Modulated symbols converter equalization conversion Discard Receiver Figure 3.11 OFDM signal generation and reception process. The mapping of the modulated data symbol into multiple subcarriers also allows an increase in the symbol duration. The symbol duration obtained through an OFDM scheme is much larger than that of a single-carrier modulation technique with a similar transmission band- width. In general, when the channel delay spread exceeds the guard time, the energy con- tained in the ISI will be much smaller with respect to the useful OFDM symbol energy, as long as the symbol duration is much larger than the channel delay spread, that is, Ts >> TRMS. Although large OFDM symbol duration is desirable to mitigate the ISI effects caused by time dispersion, large OFDM symbol duration can further reduce the ability to alleviate the effects of fast fading, particularly, if the symbol period is large compared to the channel coherence time, then the channel can no longer be considered as time-invariant over the OFDM symbol duration; therefore, this will introduce the inter-subcarrier orthogonality loss. This can affect the performance in fast fading conditions. Hence, the symbol duration should be kept smaller than the minimum channel coherence time. Since the channel 320 Chapter 3 coherence time is inversely proportional to the maximum Doppler spread, the symbol dura- tion Ts must, in general, be chosen such that << / VRMS. The large number of OFDM subcarriers makes the bandwidth of the individual subcarriers small relative to the overall signal bandwidth. With an adequate number of subcarriers, the inter-carrier spacing is much smaller than the channel coherence bandwidth. Since the chan- nel coherence bandwidth is inversely proportional to the channel delay spread TRMS, the

sub- carrier separation is generally designed such that 1/TRMS- In this case, the fading on each subcarrier is flat and can be modeled as a complex-valued constant channel gain. The individual reception of the modulated symbols transmitted on each subcarrier is therefore simplified to the case of a flat-fading channel. This enables a straightforward introduction of advanced MIMO schemes. Furthermore, in order to mitigate Doppler spread effects, the inter-carrier spacing should be much larger than the RMS Doppler spread VRMS. Since the OFDM sampling frequency is typically larger than the actual signal bandwidth, only a subset of subcarriers is used to carry modulated symbols. The remaining subcarriers are left inactive prior to the IFFT and are referred to as guard subcarriers. The split between the active and the inactive subcarriers is determined based on the spectral sharing and regula- tory constraints, such as the bandwidth allocation and the spectral mask. An OFDM trans- mitter diagram is shown in Fig. 3.11. The incoming bit stream is channel coded and modulated to form the complex-valued modulated symbols. The modulated symbols are converted from serial to parallel with N <NFFT complex-valued numbers per block, where NFFT is the size of FFT/IFFT operation. Each block is processed by an IFFT and the output of the IFFT forms an OFDM symbol, which is converted back to serial data for transmis- sion. A guard interval or CP is inserted between symbols to eliminate ISI effects caused by multipath distortion. The discrete symbols are windowed/filtered and converted to an analog signal for RF upconversion. The reverse process is performed at the receiver. A one-tap equalizer is usually used for each subcarrier to correct channel distortion. The tap coeffi- cients are calculated based on channel information. When there is multipath distortion, a conventional single-carrier wideband transmission sys- tem suffers from frequency-selective fading. A complex adaptive equalizer must be used to equalize the in-band fading

. The number of taps required for the equalizer is proportional to the symbol rate and the multipath delay. For an OFDM system, if the guard interval is larger than the multipath delay, the ISI can be eliminated and orthogonality can be main- tained among subcarriers. Since each OFDM subcarrier occupies a very narrow spectrum, in the order of a few kHz, even under severe multipath distortion, they are only subject to flat fading. In other words, the OFDM converts a wideband frequency-selective fading channel to a series of narrowband frequency non-selective fading subchannels by using the parallel mul- ticarrier transmission scheme. Since OFDM data subcarriers are statistically independent and identically distributed, based on the central-limit theorem, when the number of subcarriers NFFT is large, the OFDM signal distribution tends to be Gaussian. New Radio Access Physical Layer Aspects (Part 1) 321 3.2.1.2 Pre- and Post-processing Signal-to-Noise Ratio In an OFDM system, the SNR is a measure of channel quality and is a key factor of link-level error assessment. There are different methods for calculation of SNR in single-antenna and multi-antenna transmission systems. For single-input/single-output systems, the SNR can be viewed as the received SNR, that is, the received SNR before the detector. The post-processing SNR is often used for MIMO links and represents the SNR after combining in the receiver and measures the likelihood that a coded message is decoded successfully. In link-level simulations, the SNR r is typically calculated using the following method. Let's vector X = transmit signal where = 1, 2, ,NTX is the complex-valued transmitted symbols from the kth transmit antenna. Note that NTX is the number of transmit antennas. It can be shown that the total transmit signal power can be obtained as 07 = trace(Rx) where Rx denotes the autocorrelation matrix of the transmitted signal. The transmit power from the kth antenna is given as =E{|xx12} = 1/NTX (uniformly distributed power). If H represents the chan

nel matrix with in which NTX and NRX denote the number of transmit and receive antennas, respectively, the received signal vector y can be calculated as y = Hx + V where complex-valued Gaussian-distributed noise vec- tor V ~ C(0,021) denotes the noise vector with respect to the size of the FFT NFFT and the number of used subcarriers before the detector Nused. We define the complex-valued Gaussian-distributed noise vector n ~ C(0,021) to be noise after the FFT operation. The receive SNR before the detector is given = the SNR after the FFT operation is ivenc as can be observed that the difference between pre-FFT and post-FFT SNRs (preF/(/ Nused NFFT is always positive, imply- ing that the FFT operation suppresses the noise and enhances the SNR. 3.2.1.3 Peak-to-Average Power Ratio The peak-to-average power ratio (PAPR) for a single-carrier modulation signal depends on its constellation and the pulse-shaping filter roll-off factor. For a Gaussian distributed p-norm of matrix H is defined as In the special case when p = 2 the norm is called the Frobenius norm and forp=`ois called the maximum norm. The Frobenius norm or Hilbert-Schmidt norm of matrix H is similar, although the latter term is often reserved for the operators on a Hilbert space. In general, this norm can be defined in the following forms: wher = denotes the conjugate trans- pose of matrix H and Si are the singular values of matrix H. The Frobenius norm is further similar to the Euclidean norm on RN and is obtained from an inner product on the space of all matrices. The Frobenius norm is submultiplicative and is very useful for numerical linear algebra. 322 Chapter 3 OFDM signal, the cumulative distribution function (CDF) 8 of PAPR for 99.0%, 99.9%, and 99.99% are approximately 8.3, 10.3, and 11.8 dB, respectively. Since the OFDM signal has a high PAPR, it could be clipped in the transmitter PA, because of its limited dynamic range or nonlinearity. Higher output back-off is required to prevent performance degradation and inter- modulation products spillin

g into adjacent channels. Therefore, RF PAs should be operated in a very large linear region. Otherwise, the signal peaks leak into nonlinear region of the PA caus- ing signal distortion. This signal distortion introduces intermodulation among the subcarriers and OOB emission. Thus the PA should be operated with large power back-offs. On the other hand, this leads to very inefficient amplification and an expensive transmitter. Thus, it is highly desir- able to reduce the PAPR. In addition to inefficient operation of the PA, a high PAPR requires larger dynamic range for the receiver analog to digital converter (ADC). To reduce the PAPR, several techniques have been proposed and used such as clipping, channel coding, temporal windowing, tone reservation, 10 and tone injection. However, most of these methods are unable to achieve simultaneously a large reduction in PAPR with low complexity and without perfor- mance degradation. The PAPR & of an OFDM signal is defined as follows: In the above equation E{.} denotes the expectation operator and n is an integer. From the central-limit theorem, for large values of NFFT, the real and imaginary values of OFDM sig- nal x(t) would have Gaussian distribution. Consequently, the amplitude of the OFDM signal has a Rayleigh distribution with zero mean and a variance of NFFT times the variance of one complex sinusoid. Assuming the samples to be mutually uncorrelated, the CDF for the peak power per OFDM symbol is given by From the above equation it can be seen that large PAPR occurs only infrequently due to rel- atively large values of NFFT used in practice. The CDF of the real-valued random variable X is defined as X Fx(x) = P(X .x), Axe R, where the right- hand side represents the probability that random variable X takes on a real value less than or equal to X. The CDF of X can be defined in terms of the probability density function f(x) as F(x) = The comple- mentary CDF (CCDF), on the other hand, is defined as P(X (>x)=1-Fx(x). Since the OFDM signal has a high PAPR, it may be cl

ipped in the transmitter power amplifier, because of its limited dynamic range or nonlinearity. Higher output back-off is required to prevent BER degradation and intermodulation products spilling into adjacent channels. However, clipping of an OFDM signal has sim- ilar effect as impulse interference against which an OFDM system is inherently robust. Computer simulations show that for a coded OFDM system, clipping of 0.5% of the time results in a BER degradation of 0.2 dB. At 0.1% clipping, the degradation is less than 0.1 dB. 10 In tone reservation method, the transmitter and the receiver reserve a subset of tones or subcarriers for gener- ating PAPR reduction signals. Those reserved tones are not used for data transmission. New Radio Access Physical Layer Aspects (Part 1) 323 BPSK with -90degrees phase rotation QPSK with -90degrees phase rotation PAPR (dB) Figure 3.12 CDF of OFDM PAPR with BPSK/QPSK modulation and NFFT = 128 [30]. Fig. 3.12 shows the CDF of OFDM PAPR for BPSK and QPSK modulation assuming a 40 MHz channel bandwidth and NFFT = 128. We further applied 90 degrees phase rotation to the subcarriers in the upper 20 MHz of the channel and investigated the effect on the PAPR reduction [47]. It is shown that large PAPR values are less likely to occur with large FFT sizes as suggested earlier. The PAs have generally a nonlinear amplitude response, where the output power is saturated for large input signals. Most applications require operation in the linear region of the PA where the output power is a linear function of the input. The larger the linear operation region or alternatively the higher saturation point, the more expensive the PA. Therefore, it is imperative to reduce the PAPR of the OFDM signal before processing through the PA. The wider bandwidth of NR compared to LTE increases the PAPR of the transmitted signal and makes it harder to achieve the same power efficiency as an LTE frontend. Current esti- mates suggest that for a single PA supplying an average transmit power of 23 dBm at the antenna

, the power from the battery will be around 2.5 W, compared to around 1.8 W for current LTE UEs. The PAPR of the 5G NR signal is 3 dB higher than an equivalent LTE waveform [48], resulting in larger back-off or higher average transmit power. Another inter- esting observation is that for 5G CP-OFDM waveform using different modulations, there is no significant difference in the CCDF function- meaning higher order modulation has mini- mal impact on maximum power reduction (MPR) and PA back-off. Fig. 3.13 shows the Chapter 3 LTE QPSK SC-FDMA 20 MHz NR QPSK CP-OFDM 20 MHz NR 16QAM CP-OFDM 20 MHz NR 64QAM CP-OFDM 20 MHz NR QPSK CP-OFDM 60 MHz NR QPSK CP-OFDM 100 MHz NR QPSK CP-OFDM 200 MHz 0.001 NR QPSK CP-OFDM 400 MHz 0.0001 Peak to average power ratio (dB) LTE SC-FDMA NR DFT-s-OFDM Shaped n/2 BPSK NR BPSK DFT-s-OFDM W shaping 20 MHz NR BPSK DFT-s-OFDM NRCP-OFDM no shaping 20 MHz NR QPSK DFT-s-OFDM 7.5 dB W shaping 20 MHz NR QPSK DFT-s-OFDM no shaping 20 MHz NR QPSK CP-OFDM 0.001 20 MHz LTE QPSK SC-FDMA 20 MHz 0.0001 Peak-to-average-power ratio (dB) Figure 3.13 Comparison of CCDF of CP-OFDM and SC-FDMA PAPRs [48]. New Radio Access Physical Layer Aspects (Part 1) 325 CCDF curves of some lower PAPR options, which can be used in cell-edge areas as well as mmWave frequency bands. We can observe that DFT-spread OFDM (DFT-S-OFDM) QPSK waveform in uplink exhibits very similar PAPR as the existing LTE single-carrier frequency division multiple access (SC-FDMA) used in the uplink; however, spectrally shaped /2 - BPSK modulation can provide up to 7.5 dB PAPR reduction which can be used in sub-6 GHz and further in mmWave bands to improve uplink link budget and cover- age [48]. 3.2.1.4 Error Vector Magnitude The modulation accuracy or the permissible signal constellation fuzziness is often measured in terms of error vector magnitude (EVM) metric. In general, the EVM is defined as the square root of the ratio of the mean error vector power to the mean reference-signal power expressed as a percentage. In other words, the EVM define

s the average constellation error with respect to the farthest constellation point (i.e., the distance between the reference- signal and measured signal points in I-Q plane). In NR, the EVM measurement is conducted for all bandwidths and each NR carrier over all allocated resource blocks and downlink subframes within 10 ms measurement period. The boundaries of the EVM measurement periods are not necessarily aligned with radio frame boundaries. 3GPP defines the reference points at which the [transmitter] EVM is measured at the receiver based on which the EVM must be measured after the FFT and a zero-forcing equalizer (per subcarrier amplitude/phase correction) in the receiver [9]. The basic unit of EVM measurement is defined over one subframe in the time domain and NRB BW subcarriers (180 kHz) in the frequency domain as follows [9]: where T is the set of symbols with the considered modulation scheme being active within the subframe, F(t) is the set of subcarriers within the NRB BW subcarriers with the considered modulation scheme being active in symbol t, I(t,,f) is the ideal signal reconstructed by the measurement equipment according to the relevant transmitter model, and Z'(t,f) is the mod- ified signal under test defined as follows [9]: where z(v) is the time-domain samples of the signal under test, Ai is the sample timing dif- ference between the FFT processing window relative to the nominal timing of the ideal sig- nal, f is the RF frequency offset, o(f) is the phase response of the transmitter chain, and 326 Chapter 3 a(f) is the amplitude response of the transmitter chain. In the above equations, the basic unit of measurement is one subframe and the equalizer is calculated over 10 subframes to reduce the impact of noise on the reference symbols. The boundaries of the 10 subframes measurement periods are not necessarily aligned with radio frame boundaries. The EVM is averaged over all allocated downlink resource blocks with the considered mod- ulation scheme in the frequency domain, and a minimum of 10 downl

ink subframes. For FDD systems, the averaging in the time domain equals the 10 subframe duration of the 10 subframes measurement period from the equalizer estimation step, whereas for TDD sys- items, the averaging in the time domain can be calculated from subframes of different frames and should have a minimum of 10 subframes averaging length. where N is the number of resource blocks with the considered modulation scheme in sub- frame i and Ndl is the number of allocated downlink subframes in one frame. While the above expressions for calculation of the EVM are the same for FR1 and FR2, the para- meters are differently defined for the two frequency ranges [9]. The permissible EVM value can be estimated from the transmitter implementation margin, if the error vector is considered noise, which is added to the channel noise. The implemen- tation margin is the excess power needed to maintain the carrier to noise ratio intact, when going from an ideal to a realistic transmitter design. The EVM cannot be measured at the antenna connector but should be measured by an ideal receiver with certain carrier recovery loop bandwidth specified in percent of the symbol rate. The measured EVM includes the effects of the transmitter filter accuracy, DAC, modulator imbalances, untracked phase noise, and PA nonlinearity. As mentioned earlier, the error vector magnitude is a measure of the dif- ference between the reference waveform and the measured [transmitted] waveform. In prac - tice, before calculating the EVM, the measured waveform is corrected by the sample timing offset and RF frequency offset, then the IQ origin offset is removed from the measured wave- form. The measured waveform is further modified by selecting the absolute phase and abso- lute amplitude of the transmitter chain. 3.2.1.5 Carrier Frequency Offset An OFDM system transmits information as a series of OFDM symbols. The time-domain samples xm(n) of the mth OFDM symbol are generated by performing IDFT on the informa- tion symbols as follows [30]: New Radio Access

Physical Layer Aspects (Part 1) 327 where NFFT and NCP denote the number of data samples and CP samples, respectively. The OFDM symbol xm(n) is transmitted through a channel hm(n) and is perturbed by a Gaussian noise Zm(n). The channel hm(n) is assumed to be block-stationary, that is, it is time-invariant over each OFDM symbol. With this assumption, the output ym(n) of the channel can be represented as ym(n) = hm(n)*xm(n) + Zm(n), where hm(n)*xm(n) = hm(k)xm(k-, n) and Zm(n) is a zero-mean additive white Gaussian noise (AWGN) k=-00 with variance 02. Since the channel impulse response hm(n) is assumed to be block- stationary, the channel response does not change over each OFDM symbol; however, the channel response hm(n) may vary across different OFDM symbols; thus it is a func- tion of the OFDM symbol index m. When the receiver oscillator is not perfectly synchronized to the transmitter oscillator, there can be a carrier frequency offset = fTX - fRX between the transmitter carrier fre- quency fTx and the receiver carrier frequency fRX. Furthermore, there may be a phase offset Ao between the transmitter carrier and the receiver carrier. The mth received symbol ym(n) can be represented as ym(n) = [hm(n)*xm(n)]exp(j{2fFon + m(NFFT + NCP) |Ts + 00}) + Zm(n) where Ts is the sampling period. The carrier frequency offset can be represented relative to the subcarrier bandwidth 1/(NFFTTs) by defining the relative fre- quency offset 8CFO = AfcfoNFFTTs. The carrier frequency offset attenuates the desired sig- nal and introduces ICI, thus decreasing the SNR. The SNR of the kth subcarrier can be expressed as where ONFFT = of the SNR on the frequency offset. In the latter equation, Ph, oz and o2 denote the total average power of channel impulse response, variance of the signal, and variance of the addi- tive noise, respectively. The subcarrier index k is dropped since the SNR is the same for all subcarriers. From this SNR expression, it is clearly seen that the effect of the frequency off- set is to decrease the signal power b

y ONEET (8CFO) and to convert the decreased power to interference power. The SNR depends not only on the frequency offset ECFO, but also on the number of subcarriers; however, as NFFT increases, converges to sinc2(8CFO). Therefore, the SNR converges to SNR(8CFO) = sinc2(8CFo)Pho2/([1- as NFFT becomes increasingly large. In the above equations, the power of inter-carrier interference as a function of relative carrier frequency offset is defined as PICI(8CFO) = [1 - (8CFO)]Pho2. In practice, the subcarrier spacing is not the same among different subcarriers due to mismatched oscillators (i.e., frequency offset), Doppler shift, and timing synchronization errors, resulting in inter-carrier interference and loss of orthogonality. It can be seen that the ICI increases with the increase of the OFDM symbol duration (or alternatively decrease of subcarrier spacing) and the frequency offset. The effects of timing offset are typically less than that of the frequency offset, provided that 328 Chapter 3 the CP is sufficiently large. It can be shown that the ICI power can be calculated as a func- tion of generic Doppler power spectrum A(v) as follows [30]: where Vmax denotes the maximum Doppler frequency. We further assume that the transmit- ted signal power is normalized. It can be noted that the ICI generated as a result of carrier frequency offset is a special case of the above equation when A(f) = s(f - fCFO) in which 8(f) represents the Dirac delta function. Using classic Jakes' model of Doppler spread where the spaced-time correlation function is defined as (t) = Jo(2) in which J0(x) denotes the zeroth-order Bessel function of the first kind, the ICI power can be written as follows: which approximately gives an upper bound on the ICI power due to Doppler spread. Comparison of the power of ICI generated by carrier frequency offset and Doppler spread suggests that the ICI impairment due to the former is higher than the latter. 3.2.1.6 Phase Noise Oscillators are used in typical radio circuits to drive the mixer used for th