RaMOD improves moving object detection using raw mmWave radar data
RaMOD: Moving-Object Detection from Raw mmWave Radar Cubes under Ego Motion
RoboticsComputer Vision and Pattern Recognition
Summary
Detecting moving objects around cars is important but tricky with radar data because the car’s own movement causes confusing signals. The authors developed RaMOD, a method that figures out the car’s motion and filters out radar signals from stationary objects to better spot moving ones. Their approach processes raw radar data directly and estimates object speeds accurately, even when radar measurements wrap around. Tests show RaMOD detects moving objects more reliably and estimates their velocities more precisely than existing radar methods.
What this means in practice
- •For automotive perception teams: Enhance real-time detection and speed estimation of moving objects around vehicles using raw mmWave radar data under vehicle movement.
- •For smart city infrastructure engineers: Improve detection of moving vehicles and pedestrians from radar sensors mounted on mobile or fixed platforms subject to ego motion.
Authors
Yuan Fang, Long Zhuang, Lingzi Xing, Xijia Wei, Jingqi Jiang, Fangzhan Shi, Simon Julier, Kevin Chetty
Abstract
Millimeter-wave (mmWave) radar provides weather-robust sensing and direct Doppler measurements, yet automotive perception pipelines often reduce its output to sparse point clouds or two-dimensional projections, leaving raw range-azimuth-Doppler (RAD) cubes underutilized. Ego motion induces Doppler responses from world-static structures, while Doppler wrapping aliases radial velocities, complicating moving-object detection and velocity estimation. We present RaMOD (Radar-based Moving-Object Detection), a real-time framework for moving-object detection and velocity estimation from raw RAD cubes acquired by a vehicle-mounted single-chip radar. RaMOD explicitly estimates frame-wise ego motion and uses the estimated radar translational velocity to predict the Doppler region of world-static structures and suppress returns within it. It temporally encodes the suppressed cubes into multi-scale features to predict RAD-space boxes with class labels. For each detected object, it predicts radial and forward velocities under raw-Doppler and Doppler-unwrapped hypotheses, together with Doppler-wrapping probabilities. On our BRIDGE-Liv dataset, RaMOD outperforms all compared raw-radar methods across RAD-cube, RA, and RD detection settings, achieving 80.8% cube mAP at IoU 0.30, 15.7 percentage points above the strongest RAD-cube baseline. Velocity MAE is 1.128 m/s, with 99.87% Doppler-wrapping accuracy.