# On Learning Spatial Structure from Pre-Beamforming Per-Antenna Range-Doppler Radar Measurements

arXiv: https://arxiv.org/abs/2604.01921 · HTML: https://mlsanity.com/paper/2604.01921/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 20 of its 36 references that are in the map.
- First posted on arXiv 2 Apr 2026.

## Builds on (most cited first, 20 of 20)

- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [Occ3D: A Large-Scale 3D Occupancy Prediction Benchmark for Autonomous Driving](https://mlsanity.com/paper/2304.14365/) (2023)
- [The Oxford Radar RobotCar Dataset: A Radar Extension to the Oxford RobotCar Dataset](https://mlsanity.com/paper/1909.01300/) (2019)
- [K-Radar: 4D Radar Object Detection for Autonomous Driving in Various Weather Conditions](https://mlsanity.com/paper/2206.08171/) (2022)
- [RADIATE: A Radar Dataset for Automotive Perception in Bad Weather](https://mlsanity.com/paper/2010.09076/) (2020)
- [Boreas: A Multi-Season Autonomous Driving Dataset](https://mlsanity.com/paper/2203.10168/) (2022)
- [Raw High-Definition Radar for Multi-Task Learning](https://mlsanity.com/paper/2112.10646/) (2021)
- [RadarScenes: A Real-World Radar Point Cloud Data Set for Automotive Applications](https://mlsanity.com/paper/2104.02493/) (2021)
- [CARRADA Dataset: Camera and Automotive Radar with Range-Angle-Doppler Annotations](https://mlsanity.com/paper/2005.01456/) (2020)
- [RADDet: Range-Azimuth-Doppler based Radar Object Detection for Dynamic Road Users](https://mlsanity.com/paper/2105.00363/) (2021)
- [RODNet: A Real-Time Radar Object Detection Network Cross-Supervised by Camera-Radar Fused Object 3D Localization](https://mlsanity.com/paper/2102.05150/) (2021)
- [RadarOcc: Robust 3D Occupancy Prediction with 4D Imaging Radar](https://mlsanity.com/paper/2405.14014/) (2024)
- [RadarGNN: Transformation Invariant Graph Neural Network for Radar-based Perception](https://mlsanity.com/paper/2304.06547/) (2023)
- [Exploring Radar Data Representations in Autonomous Driving: A Comprehensive Review](https://mlsanity.com/paper/2312.04861/) (2023)
- [LiCROcc: Teach Radar for Accurate Semantic Occupancy Prediction using LiDAR and Camera](https://mlsanity.com/paper/2407.16197/) (2024)
- [Probably Unknown: Deep Inverse Sensor Modelling In Radar](https://mlsanity.com/paper/1810.08151/) (2018)
- [ADCNet: Learning from Raw Radar Data via Distillation](https://mlsanity.com/paper/2303.11420/) (2023)
- [A Deep Automotive Radar Detector using the RaDelft Dataset](https://mlsanity.com/paper/2406.04723/) (2024)
- [Towards Foundational Models for Single-Chip Radar](https://mlsanity.com/paper/2509.12482/) (2025)
- [Enhanced 3D Object Detection via Diverse Feature Representations of 4D Radar Tensor](https://mlsanity.com/paper/2502.06114/) (2025)

## Abstract (arXiv)

Automotive radar perception pipelines commonly construct angle-domain representations via beamforming before applying learning-based models. This work instead investigates a representational question: can meaningful spatial structure be learned directly from pre-beamforming per-antenna range-Doppler (RD) measurements? Experiments are conducted on a 6-TX x 8-RX (48 virtual antennas) commodity automotive radar employing an A/B chirp-sequence frequency-modulated continuous-wave (CS-FMCW) transmit scheme, in which the effective transmit aperture varies between chirps (single-TX vs multi-TX), enabling controlled analyses of chirp-dependent transmit configurations. We operate on pre-beamforming per-antenna RD tensors using a dual-chirp shared-weight encoder trained in an end-to-end, fully data-driven manner, and evaluate spatial recoverability using bird's-eye-view (BEV) occupancy as a geometric probe rather than a performance-driven objective. Supervision is visibility-aware and cross-modal, derived from LiDAR with explicit modeling of the radar field-of-view and occlusion-aware LiDAR observability via ray-based visibility. Through analyses of signal properties, transmit configurations (A-only, B-only, and A+B), receive aperture, and range-Doppler structure, together with physics-aligned baselines, we investigate the factors influencing spatial recoverability. The results indicate that meaningful spatial structure is recoverable from pre-beamforming per-antenna RD tensors under the studied A/B CS-FMCW radar configuration through learned spatial mixing, without relying on hand-crafted signal-processing stages.