Study On Signal Detection And Precoding In MIMO Communication Systems

The multiple-antenna technique introduces a new space free degree for the design of wireless communications. And this space free degree can be utilized to provide spatial multiplexing gain and diversity gain, which means that the multiple-input multiple-output (MIMO) system can increase the information rate and improve the communication quality simultaneously. When the multiple-antenna technique is utilized mainly to increase the information rate, the transmitted codeword does not support simplified signal processing at the receiver in general. Therefore, the signal detection of the MIMO system will become difficult. In this dissertation, the sequential Monte Carlo (SMC)-based MIMO detection is studied. Several methods, including noise enhancement, partially deterministic reallocation, Markov chain Monte Carlo transition and double systematic resampling, are proposed to increase the diversity of particles, thereby, improve the performance of the SMC detection. Further, two methods are presented to reduce the complexity of the SMC detection, by utilizing the sphere constraint and the multiple-level bit mapping of the signal constellation, respectively. In addition, a tree-based MIMO detection model is shown, by which most of the existing MIMO detection methods can be analyzed and compared systematically.In the adaptive coded-modulation system, the digital modulation constellation is usually not known perfectly at the receiver, so it is necessary to determine the modulation constellation before the signal detection. In this dissertation, the SMC-based digital modulation constellation classification is studied. We propose a class of SMC-based digital modulation constellation classification methods for the MIMO flat fading channel and the single antenna intersymbol interference channel, respectively. The proposed methods perform well on various constellations with different cardinalities.Concatenated with error-correcting code, the concatenated space-time code can approach the MIMO channel capacity by utilizing the joint iterative detection and decoding at the receiver. In this dissertation, the joint detection and decoding is researched. An iterative multistage decoding method is presented as the soft-input soft-output MIMO detection, and this method can achieve good detection performance with reduced complexity. Further, an iteration-stopping method for the joint iterative detection and decoding is given, which can reduce the complexity and decrease the decoding delay effectively. Moreover, an estimate of the extrinsic information clipping value is derived for the soft-input soft-output iterative tree search MIMO detection, and this method achieves better performance compared to the conventional method with constant setting of the clipping value.Motivated by the immense success in the single user communication system, the multiple-antenna technique is extended to the multiuser communication system. In this dissertation, the vector perturbation precoding for the MIMO broadcast channel is investigated. A vector perturbation precoding method maximizing the detection signal-to-interference-plus-noise ratio is proposed, which has better performance compared to the method maximizing the detection signal-to-interference-noise ratio. Moreover, with the aid of the lattice reduction and by utilizing the closest point search, a low-complexity peak-to-average power ratio reduction method is given and an effective plus amplitude modulation constellation quantization of the channel input is presented.