Interference Suppression for FMCW Radars with EMD Enhanced Sparse Matrix Recovery

This paper deals with the issue of long-time interference suppression for frequency modulated continuous wave (FMCW) radars. Conventionally, the sparse matrix recovery-based method is used to suppress the interference with the assumption that the interference is sparse enough. However, the sparse assumption cannot be fulfilled when the interference widely spreads over the time domain. We hereby propose an interference suppression method based on empirical mode decomposition (EMD) and sparse matrix recovery. Firstly, the EMD is applied to decompose the received echo signals of the FMCW radar, resulting in a series of intrinsic mode functions (IMF). The cumulative mean squared error (CMSE) is then performed to specifically reorganize the IMF components, which not only improves the initial signal-to-noise ratio, but enhances sparsity of the interference in the echo signals greatly. Then, the sparsity enhanced echo signals are transformed into a low-rank Hankel matrix and the robust principal component analysis (RPCA) method can be performed to recover the echo efficiently. Extensive simulations demonstrate the superior performance of the proposed method with long interference signal periods and strong interference intensity in comparison with state-of-the-art methods.