Research On MIMO Physical Layer Network Coding Signal Dection Based On Two Way Relay Channels

At the beginning of the 21 st Century, network coding(NC) has made a great breakthrough in the realm of information theory, which allows network nodes involved in coding, and improves the network throughput and the robustness of the system. The broadcasting nature of wireless networks makes the physical-layer network coding(PNC) become a hot topic in recent years. PNC technique can not only increase the throughput and data rate of the network, but also can reduce the power consumption and enhances the safety of the network. Multiple-input multiple-output(MIMO) technique, providing several parallel wireless links through implementing multiple antennas at both the transmitter and receiver, can significantly increase the capacity and reliability by utilizing spatial multiplexing and diversity. The aforementioned two techniques can greatly enhance the performance of the communication systems, thus we focus on the combination of them in this dissertation.First, we introduce a typical application scenario in wireless network, two-way relay channel(TWRC) model, and give concrete deta ils to PNC and mapping techniques. Then, the MIMO PNC is applied to the TWRCs. Compared to the single input single output(SISO) system, the MIMO PNC has better performance in terms of throughput and reliability, but it brings a series of problems at the same time, especially for the signal detection. The larger number of antennas, the more interference the signals will have. Therefore, how to reduce the interference between the signals and how to address the error propagation problem are two key issues to the detection algorithms.Second, the traditional MIMO PNC detection techniques consist of ZF detection, MMSE detection, and ML detection. The performance of the ZF and MMSE detections is not desirable, while the complexity of ML detection is prohibitively high. Chapter 4 proposes a new detection algorithm based on successive interference cancellation(SIC), named QR-GSIC, to improve the detection performance without a significant increase in complexity. Numerical results of the proposed algorithm are p rovided through Matlab simulation with respect to bit error rate(BER) performance.Finally, the Lenstra-Lenstra-Lovasz(LLL) can reduce the condition number of the channel matrix, thereby reducing noise interference. We apply the LLL algorithm to our proposed QR-GSIC algorithm to further improve the detection performance via suppressing the noise. The simulation results verify that the QR-GSIC LLL algorithm can further improve the detection performance compared to that of the QR-GSIC algorithm.In this dissertation, we start with MIMO and PNC, followed by TWRC as a platform. We focus on the applicability of MIMO-PNC to TWRCs and resolve the problems of signal detection of the mixed systems. Moreover, we review some of the classical detection algorithms and make tremendous improvements to them, which justified by intensive simulations.