Compared to uniform linear arrays (ULAs), uniform planar arrays (UPAs) provide a more flexible and compact deployment structure for massive multiple-input multiple-output (MIMO). Considering the large number of antennas in UPA, the conventional downlink channel estimation method has a high overhead in pilot training. To deal with this problem, we develop an off-grid sparse Bayesian learning (SBL) algorithm for downlink channel estimation based on compressed sensing (CS). Specifically, we eliminate the coupling of azimuth-angles and elevation-angles in the steering vectors by angle redefinition. Moreover, we introduce a selection strategy for grid point refinement to guarantee algorithm convergence. Simulation results reveal that the proposed off-grid SBL algorithm exhibits better performance than orthogonal matching pursuit (OMP) and existing SBL-based algorithms.
Off-Grid Channel Estimation for Uniform Planar Arrays Using Sparse Bayesian Learning
2024-06-24
387229 byte
Conference paper
Electronic Resource
English