Interference Mitigation for Automotive Radars Based on Non-convex Sparse and Low-rank Decomposition

Many potential traffic safety hazards are caused by the interference between millimeter wave radars widely used in vehicles and road equipment. To this end, a signal separation method based on non-convex regularized sparse and low-rank decomposition to mitigate the mutual interference between radars is proposed in this paper. Firstly, the low-rank and sparse characteristics of target and interference components in radar echo signals are used to construct an optimization model, and the problem of interference mitigation is transformed into an optimization problem. Then, iterative optimization is performed using the alternating direction method of multipliers under the condition of using the maximal-minimal non-convex sparse penalty. Finally, the iterative process ends within a specified error range. The target and interference components are solved separately, and the useful signals and the interference signals are accurately separated. Compared with existing signal processing methods, the proposed method skips the cumbersome interference detection step, and the correlation coefficient between the interference-suppressed signal and the pure reference signal reaches 0.999 7. The results of simulation and subsequent experiment show that the proposed method has excellent performances in interference mitigation between radars.