Two-dimensional compressed sensing of channel state information in massive MIMO system

Abstract

In most of the existing compressed sensing based Massive MIMO channel state information acquisition schemes, the huge overhead introduced by reconstruction and measurement of channel cannot afford by power-limited users. Combining channel estimation and feedback together, a Two-dimensional compressed sensing schemes is proposed for Massive MIMO channel state information. Instead of reconstruct the channel vector, Massive MIMO users compressed the spatial domain measurements from time domain and feed the measurement back to the BS and reconstruct the channel. Since the reconstruction is transferred from user side to BS, the new method alleviates the requirement of computation and memory of user side. Numerical results show that the proposed algorithm can significantly reduce the channel state information acquisition overhead with acceptable accuracy.