Multipath Signal Mitigation for Indoor Localization Based on MIMO FMCW Radar System

Indoor human localization (IHL) is an important application of the Internet of Things. Multipath effects are the most challenging problem for successful IHL when using radar, and thus, they should be mitigated properly. Therefore, effective solutions for indoor multipath mitigation are devised in this article based on a co-located multiple-input multiple-output (MIMO) frequency-modulated continuous-wave (FMCW) radar without a priori information of reflection geometry and training data sets. To this end, a velocity-azimuth domain suitable for false alarm reduction under severe multipath propagations rather than conventional domains is suggested based on an in-depth analysis of radar echoes from the indoor environment. Then, intra- and inter-frame integration are exploited to enhance the human signal-to-interference-plus-noise ratio, which helps improve the human detection probability. Finally, the detected signal is identified as a multipath based on whether the direction-of-arrival and direction-of-departure are different. The proposed framework is demonstrated through experiments conducted in a seminar room using a commercial MIMO FMCW radar. The results indicate that the multipath components are explicitly discriminated from the backscattering of an individual using the proposed scheme. Ghost targets induced by multipath can be suppresed owing to the proposed framework, and therefore robust IHL can be achieved.