Millimeter-Wave MIMO Radar for Ego-Motion Estimation and Self-Localization

Millimeter-wave MIMO radar can simultaneously measure the range, angle, and relative velocity of surrounding objects. These measurements can be used for self-localization, ego-motion estimation, and environment perception for vehicles and mobile robots. Compared with cameras and LiDAR, radar is relatively robust under low-light conditions, fog, rain, dust, and other challenging environments.

This research investigates signal processing, tracking-filter, and sensor-fusion methods for estimating surrounding objects, ego-velocity, and ego-trajectories using range, angle, and Doppler velocity information obtained from millimeter-wave MIMO radar.

Background

Accurate ego-motion estimation is essential for autonomous driving, mobile robotics, outdoor autonomous navigation, and infrastructure inspection. GNSS, IMU, cameras, and LiDAR are widely used for this purpose, but each sensor has limitations in urban multipath environments, tunnels, indoor spaces, low-light conditions, and adverse weather. Millimeter-wave radar can directly observe Doppler velocity in addition to range and angle. This velocity information is useful for ego-velocity estimation, separation of stationary and moving objects, tracking of surrounding objects, and improvement of radar odometry.

Research Objectives

This research aims to estimate self-localization and ego-motion of vehicles and robots by using multiple reflection points and surrounding objects observed by millimeter-wave MIMO radar. Main topics include:

Self-localization using millimeter-wave MIMO radar
Ego-velocity estimation using Doppler velocity
Multi-target tracking of surrounding objects
Separation of stationary and moving objects
Trajectory estimation using tracking filters such as Kalman filters
Sensor fusion with radar, IMU, cameras, LiDAR, and GNSS
Velocity ambiguity in high-resolution millimeter-wave radar

Related Work

Our previous work has investigated methods for tracking multiple reflection points and surrounding objects observed by millimeter-wave MIMO radar and estimating vehicle or robot trajectories from their temporal changes. The research focuses on multi-target tracking filters, Doppler-based ego-velocity estimation, and radar-based self-localization. It also addresses issues such as velocity ambiguity in high-resolution radar and improvement of estimation accuracy through the use of multiple sensors.

Selected Publications

Ryuto Terawake, Keiji Jimi, and Kenshi Saho, "Millimeter-wave Radar-based Ego-Trajectory Estimation of a Vehicle Using Multi-Target Tracking Filter," IEEE Sensors Letters, vol. 10, art no.3501804 , March 2026. DOI: 10.1109/LSENS.2026.3669681
Ryuei Maruyama, Ryuto Terawake, Keiji Jimi ,and Kenshi Saho, "Ego-trajectory Estimation of Electric Wheelchair Using Millimeter‑Wave Radar," CEUR Workshp Proceedings, vol-4082, pp.63-70, September 2025. DOI: 10.5281/zenodo.19583130

Keywords

millimeter-wave MIMO radar, millimeter-wave radar, FMCW radar, radar odometry, self-localization, ego-motion estimation, ego-trajectory estimation, Doppler velocity, ego-velocity estimation, multi-target tracking, Kalman filter, sensor fusion, autonomous driving, mobile robot, radar signal processing

Google Sites

Report abuse