A robust direction-of-arrival estimation method for impulsive noise environments

The performance of conventional direction-of-arrival (DOA) estimation methods degrades greatly when there are few snapshots and the noise model is mismatched, especially in impulsive noise environments. To solve this problem, this paper proposes a robust DOA estimation method, which is robust to the prob-ability density function of impulsive noise. First, by exploiting the low-rank decomposition of the residual fitting error matrix and the characteristics of impulsive noise, three conditions of a good cost function in impulsive noise environments are proposed, and a unified cost function framework is established. Based on this, a robust cost function is designed to suppress the contribution of samples with impulsive noise to the cost function. Then, a bi-iterative complex fixed-point algorithm (BI-CFPA) is developed for the nonconvex optimization problem of signal subspace estimation based on the proposed robust cost func-tion. Theoretical analyses indicate that the BI-CFPA has excellent numerical stability and low computa-tional complexity. Finally, with the estimated signal subspace, the multiple signal classification technique is employed to obtain DOA estimates. Simulation results show that the proposed DOA estimator performs better than existing methods in three typical impulsive noise environments, especially under fairly strong impulsive noise.& COPY; 2023 Elsevier B.V. All rights reserved.

Automated summary

Conventional DOA estimation methods perform poorly in impulsive noise environments with few snapshots and mismatched noise models. To address this issue, this paper proposes a robust DOA estimation method that is robust to the probability density function of impulsive noise. The method involves establishing a unified cost function framework and designing a robust cost function to suppress the contribution of samples with impulsive noise. A bi-iterative complex fixed-point algorithm is then developed for the nonconvex optimization problem. Simulation results demonstrate that the proposed method outperforms existing methods in impulsive noise environments.