Research On Signal Detection Algorithms In Multiple Input Multiple Output System

At present in the field of wireless communication, multiple input multiple output(MIMO)communication system becomes a hot area of research. Because of its property that it can increase the channel capacity and spectrum efficiency without expanding additional bandwidth and power. Since the MIMO system uses multi-antenna to transmit signals at the same time, then signals in the receiver consist of multiple tributary signals which makes the detection of received signals more difficult. Signal detection algorithm is so important for the performance of MIMO system that it is necessary to study the MIMO signal detection technology.This paper firstly focuses on the analysis of traditional detection algorithms of MIMO communication system, and simulates the linear detection algorithm and nonlinear detection algorithm based on lattice reduction algorithm. Secondly, according to the select order of candidate constellation points, this paper analyzes VB algorithm based on F-P strategy and CL algorithm based on S-E strategy. Given that noise, CL-MMSE algorithm is proposed on the basis of CL algorithm and to be simulated. Finally, in order to select the initial search radius of hypersphere exactly, a detection algorithm named as LLL-MMSE-SD based on lattice reduction preprocessing is proposed. The preprocessed channel matrix is closer to orthogonality, which makes the value of initial radius more accurate and the search radius smaller and the computational complexity lower.The simulation shows that the improved linear detection algorithm and nonlinear detection algorithm based on lattice reduction algorithm has a certain degree of improvement in capability. The CL-MMSE algorithm obtains lower computational complexity than CL algorithm at the cost of capability loss to a certain extent, which proves the algorithm is feasible. According to the actual channel correlation and ill-conditioning, the improved algorithm LLL-MMSE-SD is put forward. In the case of a little increase of bit error rate, the computation complexity is greatly reduced, which is more suitable for real-time system. Thefeasibility and effectiveness of the proposed algorithm is also verified.