Abstract:Robust perception is a vital component for ensuring safe autonomous and assisted driving. Automotive radar (77 to 81 GHz), which offers weather-resilient sensing, provides a complementary capability to the vision- or LiDAR-based autonomous driving systems. Raw radio-frequency (RF) radar tensors contain rich spatiotemporal semantics besides 3D location information. The majority of previous methods take in 3D (Doppler-range-azimuth) RF radar tensors, allowing prediction of an object's location, heading angle, and size in bird's-eye-view (BEV). However, they lack the ability to at the same time infer objects' size, orientation, and identity in the 3D space. To overcome this limitation, we propose an efficient joint architecture called CenterRadarNet, designed to facilitate high-resolution representation learning from 4D (Doppler-range-azimuth-elevation) radar data for 3D object detection and re-identification (re-ID) tasks. As a single-stage 3D object detector, CenterRadarNet directly infers the BEV object distribution confidence maps, corresponding 3D bounding box attributes, and appearance embedding for each pixel. Moreover, we build an online tracker utilizing the learned appearance embedding for re-ID. CenterRadarNet achieves the state-of-the-art result on the K-Radar 3D object detection benchmark. In addition, we present the first 3D object-tracking result using radar on the K-Radar dataset V2. In diverse driving scenarios, CenterRadarNet shows consistent, robust performance, emphasizing its wide applicability.

Target Location and Tracking of Network Radar

Radar2: Passive Spy Radar Detection and Localization Using COTS Mmwave Radar

3-Dimensional Target Localization Accuracy Analysis Based on Radar Network with Aerial and Ground-Based Nodes

Target Tracking based on 2D Radar and Infrared Sensor

The Resolution Of The Network Radar

Learning to Track Multiple Radar Targets with Long Short-Term Memory Networks

Track Initiation with Passive Electronic Reconnaissance Equipments Based on Radar Network

Adaptive Resource Management Based on Target Priority in the Radar Networks

Research of Radar Automatic Target Recognition in Practical High Resolution Range Profile

An Improved Data Association Algorithm for Networked Radars Target Tracking

CenterRadarNet: Joint 3D Object Detection and Tracking Framework using 4D FMCW Radar

Integrated Detection and Tracking in Asynchronous Moving Radar Network

Target Detection in Bistatic Radar Networks: Node Placement and Dynamic Frequency Selection

LPI based resource management for target tracking in distributed radar network

Target Location for Passive Radar in Single Frequency Network

A 3D Target Tracking Algorithm in Distributed 2D Radar Network

A Method For Improving Target Tracking Accuracy Of Vehicle Radar

Resource management for target tracking in distributed radar network system

A Track-Before-Detect Algorithm for UWB Radar Sensor Networks

Radar Selection for Single-Target Tracking in Radar Networks

Target Localization Based on High Resolution Mode of MIMO Radar with Widely Separated Antennas