Study Of Wireless Channel State Information Feedback Scheme Based On Compressed Sensing

Currently the research of the 4th generation mobile communication technology standards and architectures has completely expanded. Supporting high-speed data services and multimedia transmission services for multi-user, while providing QoS to meet the different user`s requirements is the development trend of future wireless communication. As a further evolution of the 3GPP LTE, LTE-Advanced is the focus of the current research. A number of technical measures have been used in LTE-Advanced systems to improve system performance, such as multi-antenna (diversity, multiplexing, and beamforming) technique, orthogonal frequency division multiplexing, scheduling, pre-coding, power control technology, adaptive modulation and so on. Implementation of these technologies is under the assumption of channel state information completely known at the transmitter. In the TDD wireless system, due to the channel reciprocity, the transmitter can determine the downlink channel CSI upon the CSI obtained through the uplink channel estimation. However, in FDD wireless system, the uplink and downlink channels occupy different frequencies and have reciprocity no longer, so we must feedback the CSI. In addition, with bandwidth increasing, the amount of the CSI feedback also increases. Reducing the amount of feedback will help to improve spectrum utilization, optimize system performance. It can be said that the CSI feedback restricts the performance of the system.In this paper, we study the new method of CSI feedback in frequency division duplex - multiple input multiple output-orthogonal frequency division multiplexing (FDD-MIMO-OFDM) wireless communication system. The method uses a new proposed theory so-called compressed sensing to feedback CSI. The compressed sensing technology can effectively reduce system processing complexity and the amount of data transmitted while maintaining near-perfect reconstruction of signals, so the system's feedback performance has been improved. Simulation results show that in the premise of the same amount of feedback, the new scheme improves the system performance.