Sparse Reconstruction of Chirplets for Automotive FMCW Radar Interference Mitigation

Abstract

Mutual interference in automotive radar scenarios is going to become a major concern as the density of vehicles with radar sensors in the roads increases. The present work tackles the problem tackles the problem of frequency modulated continuous wave (FMCW) to FMCW interference. In this context, we propose a signal processing technique to blindly identify and remove interference by using the fast Orthogonal Matching Pursuit (OMP) algorithm to project the interference signals in a reduced chirplet basis, and separate it from the target signal with minimal loss of information. Significant reduction of the noise-plus-interference levels are observed in measured data acquired with state of the art automotive sensors.