Discriminating Distributed Targets in Automotive Radar Using Fuzzy L-Shell Clustering Algorithm

Automotive radar is the commonly used sensor for autonomous driving and active safety. Modern automotive radars provide high spatial information on the host vehicle surroundings, and therefore, automotive radar targets appear as point clouds of radar detections. This article addresses the problem of discriminating between adjacent distributed targets using the distribution of radar detections in the range–azimuth domain. The proposed approach considers both the statistical information of the radar detections' distribution and the L-shape model of the target vehicles via the fuzzy L-shell clustering algorithm. The performance of the proposed approach is evaluated via simulations, and its superiority over the conventional methods is demonstrated in practical automotive scenarios.