| With the rapid development of mobile communication industry, relay technique was considered as an important candidate for LTE-Advanced system, which can improve the cell throughput, extend the coverage and optimize the channel quality for the cell-edge users. Therefore, relay networks have attracted significant interests from researchers all over the world. However, most of the previous works assumed full channel state information (CSI), which is not possible in reality. In order to have a deeper sight on relay system under imper-fect CSI, the thesis launched this research. Specifically, this work can be categorized as four parts:First, channel estimation error modeling for relay system. Consider the relay deployment in real world, the perfect CSI cannot be achieved easily, due to the existence of channel esti-mation errors, quantization errors and feedback delay. Apparently, having knowledge on the statistical property of non-ideal factors is of great important for relaying applications. Hence, the thesis explored the problem on channel estimation error modeling in this thesis. Based on the LTE link level simulation platform, the thesis did a lot of investigations on channel esti-mation model and took the transmitted SNR (Signal-to-Noise Ratio), user’s velocity and pilot space as effect factors. Using the Matlab software, the thesis first pointed out the incorrect assumptions which treated the imperfect CSI as a zero mean complex Gaussian variable. Then, the thesis proposed a scheme to fit the channel estimation errors by Gamma function and uni-form function. Simulation results verified our proposed model.Second, signal detection algorithm for two way relay system. Two way relay is key branch in the research of relay due to its high spectral efficiency. The thesis investigated the signal detection algorithm in two way relay system under imperfect CSI. A variable gain transceiver structure was proposed based on the training signals from both of the source nodes and relay node. During the training period, each node can estimate each link by the help of training signals and prepare for the data transmission period. Besides, the maximum likelih-ood (ML) estimator and the linear minimum mean square error (LMMSE) estimator in two way relay system were explored. Furthermore, the thesis also presented the discrepancy of these two estimators.In simulation parts, the thesis compared the characters of each estimator respectively, and pointed out the great potential of two way relay in the angle of green energy.Third, joint source-relay precoder and decoder design for single data flow. Precoding technique is one of the most important fields in modern wireless communication system. Source node can dynamically allocate the power according to the CSI feedback by the desti-nation node to achieve high throughput. There are lots of classic literatures on precoding area, but most of them assumed perfect CSI which is not convincible in real environment. There-fore, a low complexity joint precoder and decoder were proposed to maximize the received SNR under part CSI conditions. The Matlab simulation results highlight the better perfor-mance which applying our proposed schemes and also demonstrate the impact arose by the channel correlation coefficientsFourth, joint source-relay precoder and decoder design for multiple data flows. Based on the third point in this thesis, the thesis extended the scheme to multiple data flows scenario. Joint precoder and decoder were investigated which aimed to maximize the average received SNR. The thesis first derived the design pattern under full CSI, then changed the scale and made it available in part CSI condition which retains the quality of low complexity. Simula-tion results show a great improvement on the received SNR and robustness under part CSI conditions.The above are the mainly research and innovation points in this thesis, we believe the above achievements can provide a significant reference for the research on relay technique in the future. |
PDF Full Text Request