Harmonic Retrieval Joint Multiple Regression: Robust DOA Estimation for FMCW Radar in the Presence of Unknown Spatially Colored Noise

Abstract

In this letter, we present a direction-of-arrival (DOA) estimation algorithm named Harmonic Retrieval joint Multiple Regression (HRMR) for FMCW radar against unknown spatially colored noise. The HRMR algorithm focuses on the direct data domain. First, the target’s IF signal is resolved and reconstructed by the selected harmonic retrieval method. Then, we regard the signal expression of each array element as an independent form of multiple regression and further achieve the DOA by iteratively estimating the regression coefficient of the reconstructed target’s IF signal from each array element one at a time. By exploiting the existence of the same noise distribution on each element whether the noise is spatially colored or Gaussian white (i.e., temporally and spatially white), the HRMR algorithm eliminates the negative effect of the spatially colored noise and demonstrates robustness. In addition, the impact of the poles of the spatially colored noise is displayed by experiments. Furthermore, we validate that the number of saturated targets of HRMR algorithm is not restricted by the number of array elements in simulations.