Abstract:Advanced driver assistance systems (ADASs) and autonomous vehicles rely on different types of sensors, such as camera, radar, ultrasonic, and LiDAR, to sense the surrounding environment. Compared with the other types of sensors, millimeter-wave automotive radar has advantages in terms of cost and reliability under bad weather conditions (e.g., snow, rain, and fog) and does not suffer from light condition variations (e.g., darkness). Typical radar devices used in today's commercial vehicles with ADAS features produce sparse point clouds in low angular resolution with a limited number of antennas. In this article, we present a machine-learning-aided signal processing chain to suppress the radar imaging blur effect introduced by the phase migration in time-division multiplexing multiple-input multiple-output radar, to generate low-level high-resolution radar bird's-eye view (BEV) spectra with rich object's features. Compared with radar point clouds, there is no information loss in radar BEV spectra. We then propose a temporal-fusion distance-tolerant single-stage object detection network, termed as TDRadarNet, and an enhanced version, TDRadarNet+, to robustly detect vehicles in both long and short ranges on radar BEVs. We introduce a first-of-its-kind multimodel dataset, containing 14 800 frames of high-resolution low-level radar BEV spectra with synchronized stereo camera RGB images and 3-D LiDAR point clouds. Our dataset achieves 0.39-m range resolution and 1.2 degrees degree azimuth angular resolution with 100-m maximum detectable range. Moreover, we create a subdataset, the Doppler Unfolding dataset, containing 244 140 beam vectors extracted from the 3-D radar data cube. With extensive testing and evaluation, we demonstrate that our Doppler unfolding network achieves 93.46% Doppler unfolding accuracy. Compared to YOLOv7, a state-of-the-art image-based object detection network, TDRadarNet, achieves a 70.3% average precision (AP) for vehicle detection, demonstrating a 21.0% improvement; TDRadarNet+ achieves a 73.9% AP, showing a 24.6% improvement in performance.

A Two-Stage Multi-Layer Perceptron for High-Resolution DOA Estimation

Spatial Sectorized Neural Network for 2-D DOA Estimation in the Full Azimuth

Two-dimensional multi-snapshot Newtonized orthogonal matching pursuit for DOA estimation

People Flow Detection Algorithm Based on a Multiencoder-Classifier Cotraining Architecture for FMCW Radar.

Tensorized Neural Layer Decomposition for 2-D DOA Estimation

A Deep Learning Based Mutual Coupling Correction and DOA Estimation Algorithm

SDOA-Net: An Efficient Deep-Learning-Based DOA Estimation Network for Imperfect Array

SDOA-Net: An Efficient Deep Learning-Based DOA Estimation Network for Imperfect Array

Deep learning approach for high-resolution DOA estimation

DOA Estimation Method Based on Improved Deep Convolutional Neural Network

Multi-DOA estimation based on the KR image tensor and improved estimation network

Antenna Failure Resilience: Deep Learning-Enabled Robust DOA Estimation with Single Snapshot Sparse Arrays

Radar Target Detection Based on OTFS Signaling: A Deep Learning Approach

A Two-step 2-D DOA Estimation Algorithm Via Underwater Acoustic Vector-Sensor Array

CAE-CNN-Based DOA Estimation Method for Low-Elevation-Angle Target

Deep Tensor 2-D DOA Estimation for URA

Efficient Deep-Learning 4D Automotive Radar Odometry Method

An automated multi-layer perceptron discriminative neural network based on Bayesian optimization achieves high-precision one-source single-snapshot direction-of-arrival estimation

Deep-Neural-Network-Enabled Vehicle Detection Using High-Resolution Automotive Radar Imaging

Direction of Arrival Estimation Based on DNN and CNN

A Novel Tree Model-based DNN to Achieve a High-Resolution DOA Estimation via Massive MIMO receive array