Robust Measurement Matrix Design Based on Compressed Sensing for DOA Estimation

It has been well known that Massive multiple-input-multiple-output (MIMO) radar can provide an excellent performance in direction of arrival (DOA) estimation. However, the significant increasing data size will seriously reduce the computational efficiency in practical application. Although compressed measurement can reduce data size and computational complexities, improper compression will enhance the environment noise. In this paper, a robust measurement matrix is designed to reduce data size and environment noise. Different from the general compressed sensing (CS) schemes, the optimization function is established by considering the overall mutual coherence of dictionary and the energy of measurement matrix, which is more suitable for noisy environment. The optimization function is highly non-convex due to the rank shrinkage of measurement matrix. To solve this problem, an alternating minimization scheme based on matrix factorization and Principal Component Analysis (PCA) is proposed. Moreover, the structure of measurement matrix is designed for massive MIMO receiver. Furthermore, numerous results demonstrate this scheme has a better estimation performance than random measurement method and general CS schemes in the noisy environment.