Research On Limited Feedback Techniques Based On Compressed Channel In Massive MIMO Systems

Throughout the development history of the whole mobile communication system, we can learn that constantly improving data transmission rates is the basic goal for each generation mobile communication system. So the fifth generation mobile communication system(5G) who has high data transmission rates has become the research focus of the communications industry. However, 5G acquiring this advantage greatly relies on the application of the massive multiple-input multiple-output(MIMO) technology. Massive MIMO is a multiple antenna technology who configures abundant antennas at the transmitter and the receiver with the purpose of providing services for multiple users simultaneously. Massive MIMO can generate precise beam current utilizing antennas at the transmitter and then obtain the array gain by means of controlling the direction of the beam current. Thus massive MIMO can improve the system capacity without adding transmission power. Massive MIMO realizing the potential advantage is based on that base station can obtain the reliable channel state information(CSI) in the downlink. In frequency-division duplex(FDD) massive MIMO systems, base station can only obtain the CSI by means of users' feedback information because of the absence of the channel reciprocity. Hence, feedback overhead will be an important performance constraint for massive MIMO system who configures a lot of antennas. Therefore, the research focus of this thesis is the method of reducing the feedback overhead in massive MIMO system. The main research contents are summarized as follows:1. This thesis proposes a CSI feedback overhead reduction method in massive MIMO system based on spatial correlation of the channel. First of all, this method takes advantage of a condition that the nonzero elements locations in sparsified channel vectors are specified to design the optimal pilot matrix and let the pilot matrix act as the projection matrix. Then it calculates the corresponding optimal recovery matrix based on the minimum mean square error(MMSE) criteria to recover the original CSI. The experiment results and theoretical analysis show that the proposed method can effectively reduce the feedback overhead of the massive MIMO system with the premise of guaranteeing feedback accuracy.2. This thesis proposes a CSI feedback overhead reduction method in massive MIMO system based on spatial and temporal correlations of the channel. First of all, this method takes advantage of spatial correlation to design the optimal pilot matrix in each time block and let the pilot matrix act as the projection matrix. Then it calculates the corresponding optimal recovery matrix in each time block based on the MMSE criteria to recover the original CSI. This method mainly makes use of temporal correlation of the channel to subtract the CSI which has been estimated in last time block, thereby reduce the feedback information in current time block. The experiment results and theoretical analysis show that the proposed method can effectively reduce the feedback overhead of the massive MIMO system with the premise of guaranteeing feedback accuracy.