Research On Imperfect CSI Feedback Of Mimo System In High-speed Train-ground Communications

Multiple-input-multiple-output (MIMO) communication technology, which uses multiple-antennas in the transmitter and receiver and takes the advantage of multiple propagation of wireless channel, is able to greatly improve capacity and performance of wireless communication system without increment of bandwidth and transmission power. Due to the significant advantage introduced by the MIMO technology, it has been widely utilized as the core technology in4G mobile communication system, and will be adopted in high speed railway communication which has more demands on wireless services. The remarkable superiority depends on whether the transmitter can get the channel state information (CSI) and do adaptive processing. Considering the limited feedback channel in actual environment, the transmitter can’t achieve the perfect CSI. Therefore, studying limited feedback in actual scene has become as a hot topic.In high speed railway communication system, the fast time-varying wireless channel will bring some challenges in the application of MIMO limited feedback. First of all, the higher Doppler shift caused by high speed mobility will result in a weak correlation of the channel. The rapid changes of channel will cause the CSI be expired thereby impacting the system performance of limited feedback system with certain feedback delay. Secondly, with the increasing of antenna arrays and demands for high precision of CSI under high-speed scenario, the feedback quantity will grow explosively. Thus, how to effectively reduce system feedback overhead is a serious problem. Based on the above two challenges, this thesis sketches the principle of MIMO limited feedback system, focuses on the analysis of influence by imperfect CSI, and proposes two feedback mechanisms resolving the two challenges respectively. Main results are outlined as follows:1) The impact of imperfect CSI feedback in MIMO system is analyzed. With the Grassmannian quantization codebook, we derive the close-form of average symbol error rate expression including antenna number, codebook length, modulate level and transmission power. The accuracy of SER expression is verified through experimental simulation. Afterwards, the influence of the CSI feedback delay is studied, whose result shows the system performance will quickly decline, and the bit error rate approaches to a platform in the high mobility scene.2) When there is feedback delay, the performance of MIMO system would be degraded with the increase of mobility speed An efficient limited feedback scheme is proposed after the investigating into various adaptive filtering prediction algorithms, the predict CSI of next transmission time based on the Kalman filter, and the transmitter modify portion of the CSI with the help of train control information. Consequently, the proposed scheme solves the above problems effectively, and makes the MIMO technology limited feedback be well suitable for high speed railway communication.3) For the secondly challenge under high speed, two limited feedback approaches are proposed. According to the temporal correlation, the variety of channel can be modeled as a finite-state Markov process. Based on the Markov model and the idea of variable-length coding, feedback reduction algorithm1without loss in system performance is derived, which can achieve the transition probability matrix for each feedback interval by fingerprint information. Algorithm2sets thresholds on the basis of algorithm1to reduce the size of codebook, and gains substantial amount of feedback reduction in exchange for a certain loss of performance. The algorithm2can remarkably reduce the feedback overhead or improve the performance with the same feedback rate compared with neighborhood feedback method and traditional feedback method.