born5149/autonomous-vehicle-sensor-fusion

Autonomous Vehicle Raw Sensor Telemetry Dataset

This repository contains raw, uncompressed sensor buffers, neural network weights, and memory dumps collected from autonomous vehicle test platforms. Data is provided "as is" for offline perception model training, system debugging, and Hardware-in-the-Loop (HIL) simulations.

Dataset Structure (Full Schema)

Core Memory & State Logic:

• autonomous_core_memory_* - Core autonomous driving kernel crash dumps and critical state memory snapshots.
• sensor_fusion_init_dump_* - Initial state vectors and calibration matrices loaded during cold boot.
• rosbag2_raw_system_telemetry_* - High-frequency ROS2 system telemetry streams and asynchronous event logs.
• ros2_middleware_trace_* - Latency and execution trace logs from the ROS2 middleware communication layer.

Spatial & Navigation Systems:

• navigation_hd_map_cache_* - Pre-compiled High-Definition (HD) route map caches and localization tiles.
• lidar_128ch_raw_spin_* - Raw 128-channel Lidar spin returns and intensity values.
• pointcloud_accumulation_buffer_* - Aggregated 3D spatial returns for dense pointcloud generation.
• occupancy_grid_3d_volume_* - 3D voxel occupancy grids for volumetric spatial mapping.

Optical & Vision Sensors:

• surround_cam_uncompressed_* - Raw, uncompressed optical feeds (.raw) from the 360-degree camera matrix.
• fisheye_surround_raw_* - Distorted, raw frame sequences from the parking camera matrix (.raw).
• stereo_depth_map_* - High-resolution, unprocessed disparity maps from the forward stereo vision system.
• night_vision_raw_feed_* - Uncompressed infrared optical streams for low-light environment perception.
• thermal_ir_array_* - Uncalibrated 16-bit thermal sensor matrices for pedestrian and animal detection.

Simulation & Machine Learning Weights:

• transformer_attention_weights_* - Massive matrix weights and intermediate states from the perception Transformer models.
• carla_unreal_engine_render_cache_* - Heavy, uncompressed rendering caches from Carla (Unreal Engine 5) simulations.
• perception_tensor_activations_* - Intermediate tensor activation states for deep learning model debugging.
• synthetic_sim_environment_* - Virtual environment state captures and simulated object behaviors.

Acoustic & Network Telemetry:

• telemetry_baseband_dump_* - Raw 5G/V2X baseband radio frequency (RF) dumps for network latency analysis.

---

[RESERVED / PENDING INGESTION PIPELINE]

The following structures are currently buffered on local ingest servers and are scheduled for upcoming synchronization:

• sensor_fusion_matrix_* - Multidimensional state correlation matrices from the centralized perception engine.
• hyperspectral_scan_* - Multi-band hyperspectral imaging returns for advanced material classification.
• bev_features_tensor_cache_* - Bird's Eye View (BEV) feature maps extracted from multi-camera rigs.
• neural_planner_replay_buffer_* - Action-state pairs for reinforcement learning and trajectory planning.
• model_checkpoint_epoch_raw_* - Uncompressed epoch checkpoints during continuous model training.
• front_telephoto_raw_stream_* - Long-range optical feeds (.raw) for high-speed object detection.
• hdr_bracket_sequence_* - High Dynamic Range exposure sequences for low-light sensor evaluation.
• stereo_disparity_volume_* - 3D disparity volumes calculated from stereo baseline calibration.
• solid_state_lidar_matrix_* - Non-mechanical solid-state Lidar returns for short-range obstacle mapping.
• pcap_ethernet_backbone_raw_* - Raw PCAP packet captures from the vehicle's internal gigabit ethernet backbone.
• nvidia_orin_gpu_memory_dump_* - Complete VRAM snapshots from the primary NVIDIA Orin compute nodes.
• nvme_swap_cache_partition_* - High-speed NVMe swap partition dumps for memory leak analysis.
• kernel_panic_crashdump_full_* - Complete system state dumps triggered during catastrophic OS kernel failures.

---

Data Collection Status & Changelog

Current Phase: Phase 2 (Heavy Simulation & High-Density Traffic Logging) System Status: [ACTIVE] Continuous data ingestion in progress.

• [Recent Updates]: Phased out legacy unstructured ECU dumps. Transitioned to high-fidelity continuous Lidar (128-channel) and heavy Unreal Engine 5 simulation caches to improve offline training accuracy.
• [Architecture Note]: To accommodate horizontal scaling, files are strictly indexed by sensor hardware or software channel (chX). Extremely large dumps or cumulative data batches (exceeding 10-20GB) generated during high-density traffic events are sharded into sequential volumes (_part1, _part2) to prevent network timeout during uploads and ensure data integrity.
• [Upload Pipeline Queue]: Advanced sensor categories (listed in the Reserved section) will be synced to this repository over the next several weeks to comply with bandwidth allocation limits.

Note to engineering team: File sizes vary significantly — ranging from small initialization matrices (~10MB) to massive raw simulation chunks, audio accumulation batches, and transformer weights (up to 50GB+). Do not truncate or compress these files during local parsing to preserve cryptographically verified entropy.