Compressed Sensing-based Channel Estimation For Fdd Multi-user Massive MIMO Systems

Massive multiple-input multiple-output(MIMO)can support high spectrum and high energy efficiency.And it is considered to be one of the key technologies to meet the needs of capacity of the next-generation wireless communication.In order to take advantage of the large-scale antenna of the base station(BS),the acquisition of accurate channel state information at the transmitter(CSIT)is indispensable.Conventional channel estimation methods require that the pilot length be at least the same as the number of transmit antennas,which will cause large time-frequency resource consumption in massive MIMO systems.Early work addressed this challenge by adopting the time division duplex(TDD)model,in which CSIT can be obtained through channel reciprocity.Compared to TDD systems,frequency division duplex(FDD)systems can provide more efficient communication with low latency,which are dominated in contemporary cellular networks.Therefore,it is necessary to study the acquisition of CSIT under the FDD systems.Many studies have shown that the massive MIMO channel has approximate sparsity.Based on the fact,the thesis mainly studies FDD massive MIMO channel estimation based on compressed sensing(CS).This thesis studies firstly the channel estimation scheme of the joint of the downlink pilot training and the uplink feedback.That is,the user equipment does not perform channel estimation after receiving the training pilot,but directly feeds the training pilot back to the BS and channel estimation is performed at the BS.Then the thesis studies the channel estimation under the point-to-point massive MIMO system.Since the channel links are temporarily correlated,previous channel support can be utilized to improve the accuracy of the current channel support identification.An improved channel estimation algorithm is proposed,and the simulation results indicate that the proposed algorithm achieves a large performance improvement.Finally,this thesis focuses on massive MIMO channel estimation under multi-user systems.According to the spatial correlation between different channel matrices,a channel matrix splitting scheme is proposed,which utilizes the characteristics of partial common support between different channel matrices by splitting the channel matrix into two more sparse channels.The simulation results demonstrate that the proposed scheme can effectively reduce the pilot overhead required for channel estimation and ensure better channel estimation performance.