Compressive Sensing and Prior Support Based Adaptive Channel Estimation in Massive MIMO

While massive MIMO obtains a great advantage over traditional MIMO, the overwhelming pilot overhead becomes a challenging problem in channel estimation. Different schemes have been presented and Compressive Sensing (CS) has been applied in this field. In this paper, the FDD massive MIMO system with temporal correlation is considered. We propose a method to compute the number of common indices between two correlated channel supports, which can be used to enhance the performance of channel estimation. Based on the number of common indices, we put forward the modified sparsity adaptive matching pursuit (M-SAMP) algorithm to exploit the prior support adaptively without sparse degree. Simulation results show that the proposed algorithm outperforms the conventional algorithms which exploit the prior support blindly.