K-Means Clustering based Sparse Cross-Entropy Minimization algorithm for DOA estimation
Qiang Guo, Hanyu Jiang, Jianhong Xiang, Yu Zhong · IET conference proceedings. · 2024
Aiming at the problem that traditional greedy class compressed sensing algorithms are influenced by noise in the estimation of direction of arrival (DOA), this paper proposes the K-Means Clustering based Sparse Cross-Entropy Minimization (KMCSCEM) algorithm. First the algorithm selects atoms by minimizing the sparse cross-entropy to improve the accuracy of atom selection at the low signal-to-noise ratio (SNR). Subsequently, K-mean clustering is added in each iteration to reduce the range of atom selection, which further improves the estimation accuracy. Finally, the speed of convergence is improved by introducing an orthogonal least squares (OLS) criterion. Simulation experiments show that the proposed algorithm has an exact estimation ratio close to 90% when the SNR is less than -5 dB.