Doppler-Assisted Automotive Radar Data Clustering Under Micro-Doppler Effect

Modern radars have high resolution and produce multiple measurements from each target in a measurement cycle. Clustering radar measurements is a significant and challenging problem. Utilizing Doppler information is an effective method to improve clustering accuracy, while the presence of the microDoppler effect could significantly affect the clustering accuracy. In this paper, a two-step clustering algorithm is proposed to group micro-Doppler measurements with those from the target body, accounting for tire position constraints and maximizing range rate likelihood. Matlab simulated results demonstrate that our proposed method significantly reduces computation time by 99 % while improving overall clustering accuracy by 2 % compared to our previous clustering approach.