Abstract: A novel signal processing algorithm is proposed for azimuth estimation of unknown targets in massive multiple-input multiple-output (MIMO) systems. The algorithm exploits the low-rank property of the received signal matrix and a hierarchical search strategy, combined with random matrix approximation, to significantly reduce computational complexity while maintaining high estimation accuracy. By applying low-rank matrix approximation, the algorithm reduces the dimensionality of the data matrix, and the hierarchical search optimizes within a limited space, avoiding the high cost of exhaustive searches. Random matrix approximation further compresses data, effectively lowering computational load. Simulation results show that the proposed algorithm reduces computational complexity by two orders of magnitude while achieving the same accuracy as classical subspace methods. This makes it well-suited for real-time azimuth estimation in large-scale MIMO systems, demonstrating its efficiency and feasibility in practical applications.

Key words: massive MIMO, randomized matrix approximation, azimuth estimation, low complexity