The Research On Signal Detection Technology In Multiple Antennas Wireless Communications

Information theoretic results show that multiple input multiple output (MIMO) system is able to offer extremely high spectrum efficiency in rich multi-path scattering environment. The MIMO capacity, which is the upper bound of information rate, can't be realized easily, because the optimal joint space time processing is impractical for high complexity. So it is important to design signal transmission form in the transmitter and detection algorithms in the receiver.The main contents for this thesis are about the detection algorithms of MIMO systems, as follows:1. Proposes two sorted QR decomposition detection algorithms based on Householder transform and Givens rotation transform. The results show that the proposed algorithms achieve the same performance of the sorted QR decomposition detection algorithm based on Gram-Schimidt transform proposed by D.Wubben. Although the detection orders obtained by the two algorithms are not optimal, they avoid matrix inversion operation, which reduce computation complexity and have little performance loss compared with V-BLAST interference cancellation algorithm. Also linear and nonlinear algorithms combined with lattice reduction are investigated and proved to improve system performance.2. Researches semidefinite relaxation(SDR) stack and branch and bound(BB) algorithms that can approach ML performance. SDR algorithm is applied to the decoding of multi-layer space time code. Modifying the visit form of nodes in Openlist of BB algorithm, we propose the minimum lower bound method to visit Openlist nodes. The sorted QR decomposition algorithms proposed in this paper are used to sort signal in BB and stack algorithms. The results show that BB and stack algorithms achieve ML performance and have better performance than SDR algorithm, but SDR algorithm approaches ML performance in polynomial complexity. The average search complexity of BB algorithm is lower than that of the stack algorithm. The improved BB and stack algorithms accelerate convergence rate, reduce computation complexity and storage space.3. Proposes an iterative BB algorithm based on list, which uses not only Openlist nodes stack but also another nodes stack called Nodelist. During...