Research On The Preprocessing And Detection Technologies For Multiple Input Multiple Output (MIMO) Wireless Communications

With the increase in the number of subscribers and the growing demand for wireless communication services, neither conventional single-input and single-output (SISO) wireless communication system nor antenna diversity technologies can meet the requirements of high spectral efficiency and link quality. Multiple-input and multiple-output (MIMO) systems combined with space-time processing technology, employing multiple antennas at both sides, are capable of exploiting the spatial properties of the multipath channel effectively. It can provide higher spectral efficiencies and enhance communication performance in rich scattering environments. MIMO is considered to be one of the key techniques that can realize high-speed wireless communication in the future.However, there are some problems for MIMO systems to be implemented relative to SISO systems. For example, spatial correlation under practical environments has reduced system capacity and reliability of MIMO. Heavily correlated channel will severely degrade the performance. On the other hand, the deployment of multiple antennas will result in higher implementation cost and detection complexity. The key factor for the application of MIMO system is how to overcome the channel correlation and reduce the algorithm complexity while retaining a certain performance level.Adaptive preprocessing has been proposed as an effective way to enhance the capacity and performance of MIMO systems in fading and correlated channels. In addition, both antenna selection technique and parallel multistage detection are efficient ways to reduce cost and complexity of MIMO systems, and enhance its practicability.This dissertation is focused on the above issues in implementing wireless MIMO systems. Various preprocessing technologies such as power allocation, beamforming,and transmit antenna selection are investigated. Furthermore, multistage parallel detection algorithms for MIMO receivers are analyzed in detail. In addition, effects of array configuration on the performance of MIMO in correlation are studied.At first, the preprocessing schemes for spatial multiplexing(SM) system to minimize error rate are studied. A power allocation stategy is proposed for the decision feedback receiver in quasi-static fading channel based on the minimum instantaneous Bit Error Rate(BER) criterion. Then, a hybrid precoding scheme combining eigen-beamforming (E-BF) and power allocation is proposed to minimize the average Symbol Error Rate(SER) by exploiting second order statistical channel state information (CSI) in correlated Rayleigh fading channel. The analytical expression for power allocation and eigen-beamforming are derived. It avoids loop or iteration operation. A similar preprocessing approach is obtained for Rician channel by equating the effect of LOS component on the MIMO performance with that of transmit correlation. And also a simple precoder only based on K factor and LOS departure angle in independent Rician channel is presented, which needs less feedback overhead and lower computation complexity.Next, the BER performance of STBC system combined with E-BF in correlated channels is studied and an antenna selection scheme based on channel statistical information is proposed. An antenna selection criterion to minimize average BER is proposed, which utilizes the same channel information as the original hybrid system and does not necessitate frequent updates. It is helpful for the hardware implementation and information feedback.Moreover, the effects of array configuration on the performance of MIMO system in correlation channels are discussed. First the relationship between Signal Noise Ratios(SNR) of every branch and average BER of the STBC combined with E-BF system is analysed. Then a nonuniform linear array (NLA) is employed in place of uniform linear array (ULA) to enhance BER performance. With a centrosymmetric structure, the uniform circular array (UCA) is also employed in the SM system to increase the system stability. The simulations have confirmed the evident advantages of UCA over ULA in highly correlated channels.Finally, the parallel multistage detection schemes used to reduce MIMO receiver complexity are investigated. In order to mitigate the error propagation, two novel multistage detection algorithms are proposed. One is an improved adaptive partial multistage detection algorithm. During every stage, the substreams are detected in an ascending order of error performance to provide more accurate cancellations. Much better performance is achieved over the traditional one especially when the number of substreams is large. Secondly, a Bi-directional parallel multistage detection algorithm is proposed. The substreams are detected in ascending and descending order simultaneously in every stage, and then the two soft decisions are combined to obtain the final estimation. Without the demand of error probability computation and optimal ordering, it has a reduced computional complexity.