Multipath Ghost Recognition and Joint Target Tracking With Wall Estimation for Indoor MIMO Radar

Abstract

In radar-based tracking systems, the presence of false detections caused by multipath effects (so-called 'ghosts') leads to the generation of false tracks and complicates the process of data association. This paper presents a novel approach for joint target tracking with wall estimation based on a robust multipath recognition and mitigation algorithm for colocated multiple-input-multiple-output (MIMO) radars. Indoor real-world radar measurements are used to recognize multipath ghosts, localize the targets, and map the wall reflectors in the room without prior knowledge of the multipath geometry, such as room boundaries. An innovative track-to-track fusion is introduced to fuse the walls estimated by multiple joint trackers. Compared to state-of-the-art methods, our approach substantially reduces the complexity of data association in scenarios with multiple targets and walls.