Research On Soft-output Signal Detection And Parameters Design Of Wireless MIMO Transmission Link

In a wireless multiple-input multiple-output (MIMO) transmission link, multiple antennas are employed at both transmitter and receiver to perform data transmission so that multiplexing and diversity gains can be simultaneously obtained. Then, the spectral and power efficiency of the wireless communications system can be improved. In order to fully utilize the advantage introduced by MIMO, signal detection algorithms should be designed to match the MIMO channel state information and make a good tradeoff between performance and complexity. Furthermore, the MIMO transmitting signals and system parameters should be optimized with respect to specific application scenarios.In this dissertation, signal detection, transmitting signal and system parameter design of wireless MIMO transmission link are investigated. For channel coded wireless MIMO transmission link, several novel soft-output MIMO signal detection schemes are proposed, which include soft-output linear minimum mean square error (MMSE) and ordering successive interference cancellation (OSIC) MMSE detectors under imperfect channel estimation, a reduced-complexity MMSE soft interference cancellation (SIC) detection algorithm applied in MIMO Turbo receiver. For the transmitting signal and system parameter design, two schemes are investigated. The first one is a near-optimum power allocation between pilot and data symbols for a wireless MIMO transmission link with linear detector. Another scheme is the optimal parameter selection of Alamouti coded orthogonal frequency division multiplexing (OFDM) system under time-varying multipath fading channels.Soft-output linear MMSE MIMO detection algorithm is firstly investigated under block-fading frequency flat channel and imperfect channel estimation. Based on the theorem of random vector and matrix, the statistics of maximum likelihood (ML) and MMSE MIMO channel estimations are analyzed under correlated MIMO channels. By using the obtained statistics of MIMO channel estimation, MMSE filter and the computation of soft information of coded bit are derived by taking both channel estimation errors and MIMO channel correlation into account. The computational complexity of the proposed schemes is also analyzed. Performance simulations show the proposed algorithms outperform the existing one with sizable gain.Following the idea of chapter 2, soft-output OSIC MMSE MIMO detection algorithm is considered in chapter 3. In order to improve the reliability of soft information, a soft-output OSIC MMSE MIMO detector based a posteriori symbol probability estimation is proposed by taking consideration of residual interference cancellation errors. A complexity-reduced implementation is also provided. Based on the proposed soft-output OSIC MMSE MIMO detector and the statistics of MIMO channel estimation obtained in Chapter 2, a novel soft-output OSIC MMSE MIMO detector is further derived under imperfect channel estimation and channel correlation. The computational complexity of this novel detector is also analyzed. Compared with the existing ones, performance simulations show significant advantage of the proposed schemes.According to the results of chapter 2 and chapter 3, chapter 4 deals with the issue of power allocation between pilot and data symbols for a wireless MIMO transmission link with linear detectors and MMSE channel estimation. For zero-forcing (ZF) detector, a soft-output implementation under channel estimation errors is derived. Based on this new ZF detector, average post-processing signal-to-noise (SNR) is analyzed. Then, aiming at maximizing the average post-processing SNR, an optimal power allocation between pilot and data symbols is presented. The performance bound of this proposed power allocation is also analyzed. For MMSE detector, the optimal power allocation is derived with respect to maximizing the average lower bound of minimum post-processing SNR. For both of these two power allocation schemes, the existence and uniqueness of the optimal value are proved. Simulation results verify that the proposed schemes can achieve near-optimum performance.In chapter 5, a complexity-reduced MMSE SIC detection algorithm in MIMO Turbo receiver is proposed. By studying the existing detection algorithms, this new MMSE SIC detection scheme is based on the reliability of the estimated a priori symbol probability. Due to the fact that the estimated a priori symbol probability can reach its real value with high probability as iteration increases, the new MMSE SIC detection algorithm can make a good tradeoff between complexity and performance.For wireless MIMO transmission link, another important issue is the impact of time-varying characteristics of multipath fading channel. This problem is also explored by focusing on the Alamouti coded OFDM systems in Chapter 6 of this dissertation. Based on subcarriers clustering and Gaussian approximation, a tight upper bound of the outage information rate of Alamouti coded OFDM systems is derived. By using this tight upper bound, the optimal selection of the number of OFDM subcarrier is provided to achieve maximal outage information rate.Finally, the dissertation is summarized in chapter 7 by reviewing the above research works. Some potential research directions in the future are also presented.