Research On MIMO Detection And OFDM Preprocessing Algorithms

Multiple-input multiple-output(MIMO) and orthogonal frequency-division multiplexing(OFDM) were considered as a key technique of the new generation of mobile communication system in the physical layer. MIMO system can significantly improve the efficiency of the frequency and the power, with no extra resource. The OFDM system has an advantage of reducing the frequency selected fading resulted from multiple paths over the single carrier system in the wireless communication system.The point of a great importance in this thesis is the signal detection technology in Vertical Bell Labs layered Space-Time(V-BLAST) system. This thesis gives an introduction to several general detection algorithms, such as linear detection algorithm(Zero-Forcing(ZF) and Minimum Mean Square Error(MMSE)), Maximum Likelihood(ML) solution and suboptimal low complexity iterative detection algorithm(including Probabilistic Data Association(PDA) and Minimum Mean Square Error-Iterative Soft Decision Interference Cancellation(MMSE-ISDIC)). Linear detection algorithm has a low complexity, but a poor performance. However ML solution has too high complexity to compare with other algorithms. This thesis makes an analysis on a low complexity iterative detection algorithm, followed by the principle of PDA detection algorithm, as well as the real Probability Density Function(PDF) of the noise and interference in it. The variance of the MMSE-ISDIC algorithm is not accuracy in noise and interference used in the algorithm. An improvement that has more accuracy description is proposed in this thesis, and it can improve the performance of the original algorithm to a certain extentAs a consequence of the time-varying channel, the orthogonality between different subcarriers is destroyed in conventional frequency-domain approaches, resulting in ICI, which increases an irreducible error floor in proportion to the normalized Doppler frequency. This thesis mainly researches on the signal detection algorithm in doubly selective channels, including the traditional detection algorithms and the low-complexity detection algorithm. Then this thesis makes an analysis of the effect of the double selective fading channel on OFDM symbols, as well as the reason of ICI, followed by the system model of OFDM in this kind of channel. Some low-complexity however will result in some performance loss. For instance, the FIR-MMSE-SICdetection algorithm neglects the channel matrix band energy, with respect to the feature of the channel matrix in double selected fading channel for OFDM. By preprocessing transmission, we propose a novel strategy which squeezes ICI into a few coefficients resulting in a better performance in low-complexity detection algorithm. By combining the transmit processing with one at receiver, the most of channel matrix can be set in a certain size band. Consequently, the loss in performance will be not so large that the power out of band for channel matrix can be ignored. The novel extension outperforms the approaches above significantly.