An Enhanced Matrix Completion-Based Interference Mitigation Method for FMCW Radar

Abstract

Translate Article

This paper addresses the challenge of cross-interference in frequency-modulated continuous-wave (FMCW) radar for autonomous driving applications. An enhanced matrix completion-based interference mitigation method for FMCW radar is proposed. The method first detects and removes the interference segments in the dechirped beat signal and reconstructs the signal into a Hankel matrix. Then, an inertia-accelerated strategy is incorporated to achieve signal recovery via matrix completion. Extensive validation using both simulated and measured data demonstrates that the proposed method can accurately reconstruct the useful signal, effectively suppress the elevated noise floor in the range spectrum, and preserve weak target information, thereby significantly improving the reliability of automotive radar under interference conditions.