Signal detection of soted QR decomposition based on log-likelihood ratio for V-BLAST OFDM systems

Abstract

High data rate and mobility for ubiquitous multimedia services are required for the next-generation wireless communication systems. As one of the most promising technologies to satisfy these requirements, V-BLAST OFDM systems are devised for high throughput and the robustness against multipath fading. In this thesis, we propose novel detection algorithms using Log-Likelihood Ratio (LLR) to detect signals for V-BLAST OFDM systems. These algorithms utilize SQRD (Sorted QR Decomposition) of the channel matrix, and apply LLR to determine the order of detection. Simulation results show that the proposed algorithms using LLR-based SQRDs provide the better performance than the conventional algorithms using SNR-based SQRDs with small overhead for additional computations. Specifically, when the number of transmit antennas $\It{n}_{T}$ is 8, and the number of receiver antennas $\It{n}_{R}$ is 8, the aver-age BER performance of our proposed algorithm using LLR-based ZF-SQRD is better than that of the conventional algorithm using SNR-based ZF-SQRD by about 2dB for QPSK. Furthermore, the average BER performance of the proposed algorithm using LLR-based MMSE-SQRD outperforms the conventional algorithm using SNR-based MMSE-SQRD by about 4dB for QPSK at $10^{-4}$ target BER. Therefore, our proposed algorithms using LLR-based SQRDs can be presented as the detection algorithm for V-BLAST OFDM systems.