Research On Pre-Processing And Frequency Domain Equalization For MIMO Wireless Systems

Multiple-input multiple-output (MIMO) systems, in which both the transmitter and the receiver are equipped with multiple antennas, can not only exploit the multi-path fading, but also improve the spectral efficiency. It has become a promising technique in the future high data-rate wireless communication system. When orthogonal frequency division multiplexing (OFDM) or single carrier frequency domain equalization (SC-FDE) is combined with MIMO, inter-symbol interference can be combated effectively. Moreover, when the channel state information is available at the transmitter, pre-processing the transmit signals will increase system throughout and enhance the reliability. In this thesis, we would study the various pre-processing techniques and frequency domain equalization for MIMO systems.To begin with, according to the minimum bit error rate (BER) criterion, we analyze the bit allocations in the MIMO multiplexing systems, i.e. vertical Bell Laboratories Layered Space-Time (V-BLAST) architecture. The transmit power is equally allocated to each antenna while the modulation mode per antenna is optimized to improve the BER performance. We proposed a greedy bit allocation and a Bisection method based bit allocation algorithm. In contrast with the conventional V-BLAST system, bit allocation per antenna is capable of improving system performance obviously. On the other hand, compared with the existing transmit power allocation scheme, bit allocation has no performance loss while it has reduced the linear range of the power amplifier and saved the cost.Second, we have studied the linear precoding technique for the downlink of the multiuser MIMO systems. Based on the analyses of the existing methods, we propose a simple joint precoding and dynamic power allocation scheme. The complex problem, which jointly optimizes linear precoding and power allocation, is simplified and decomposed into two separate steps. On the basis of the precoding, transmit power for different users is optimized to make the signal-interference-noise ratio (SINR) of each user equal and maximum, with the result that the BER performance is enhanced. In addition, we have simplified the existing minimum mean square error (MMSE) linear precoding and proposed a suboptimum MMSE (Sub-MMSE) pecoding method. By introducing a scalar gain to the decoding vector for each user, we present a greedy algorithm to obtain a suboptimum solution to the normalized decoding vectors. Then, the transmitter precoding vectors and the scalar gain are optimized based on MMSE criterion. The performance of the proposed Sub-MMSE method can approach that of the optimum MMSE scheme with lower complexity.Moreover, the nonlinear precoding for the downlink multiuser MIMO channels is investigated. A Tomlinson-Harashima precoding (THP) scheme is proposed for multiuser MIMO system with multiple antennas at each receiver. Assuming single data stream communication for each user, joint transmitter and receiver design is done to maximize the signal to noise ratio (SNR) for each user. A heuristic user ordering algorithm is proposed to optimize the encoding order and improve the bit error rate (BER) performance with acceptable complexity. Then, we present a multiuser MIMO broadcast scheme with dirty paper coding (DPC) at the transmitter and linear equalization at the receiver. A multiuser scheduling algorithm is presented to exploit multiuser diversity when the number of the users is larger than that of transmit antennas. The proposed scheme can achieve the sum rate close to the Sato bound and superior to some of the existing schemes. Furthermore, the scheme is extended to the multiuser MIMO-OFDM downlink. At each subcarrier, every user utilizes the left-singular vectors of its channel matrix as receiver. DPC precoding is applied to allocate the spatial subchannels to users. Water-filling power allocation is performed to achieve the optimum sum-rate capacity. The proposed method could approach the Sato bound with lower complexity and provide a tradeoff between system performance and partial feedback.Finally, we present the frequency domain equalization for single carrier single-input single- output system. Based on the zero forcing (ZF) and MMSE criteria respectively, we derived two noise-predictive equalization schemes for unique-word (UW) based single-carrier systems. The correlation properties of the noises in the outputs of the frequency domain equalizer have been exploited to predict and cancel the noises in the estimation of data. The proposal is further extended to MIMO single-carrier systems. Theoretical analyses show that both of the proposed techniques perform better than the conventional frequency domain equalizers. Simulation results have confirmed the significant performance improvement they can achieve.