A Study on Object Detection and Tracking by Millimeter Wave FMCW Radar

Abstract

This paper presents a complete FMCW radar based target tracking system using the TI AWR1843 mmWave sensor and DCA1000EVM. Real radar data captured through mmWave Studio, along with synthetically generated FMCW signals, are processed in MATLAB for controlled evaluation. A walking human target is tracked using range, Doppler, and angle estimation, followed by alpha-beta and Kalman filtering to obtain position, velocity, and trajectory. Performance metrics such as range RMSE, angle RMSE, Cartesian ARMSE, and radial-velocity RMSE are used to quantify accuracy. Results from both real and synthetic scenarios show reliable 1D and 2D tracking, demonstrating the effectiveness of the proposed radar processing and tracking methodology for short-range human motion analysis.