Research On MIMO Signal Detection Algorithms In LTE-A System

MIMO wireless communication technology provides a lot of potential development space toimprove spectrum efficiency and improve the transmission performance of the wireless link. It isthis feature that makes the MIMO technology become one of the core technology of next generationcellular communication system. Around the research on anti-interference receiver antenna signaldetection algorithm of the next generation mobile communication technology LTE-A system isalso a research focus of many experts, scholars and engineers. Signal detection is essentially theprocess of using received vectors Y at the receiving end and channel matrix H, which are obtainedby channel estimation,to estimate the sending vector X based on different signal detectionalgorithm.Multi-input Multi-output (MIMO) signal detection algorithm often should have a goodapproximate maximum likelihood detection performance, and has lower computational complexity.However, both in the actual situation is often impossible to have it all.This paper first briefly introduces the new characteristics and new technologies in the LTE-Asystem, then introduces the basic theoretical knowledge of MIMO wireless communication system,and the application of MIMO in LTE-A system, followed by the study of the existing classicalMIMO signal detection algorithms, and then compares their signal detection performance andcomputational complexity by the computer simulation.Finally, also the innovation of this paper, inview of establishing the LTE-A single frequency reuse cellular downlink model, puts forward asphere decoding algorithm which has fixed-complexity step by step,but there are still manyshortcomings of the algorithm. Then according to the highlighted sphere decoding problem, furtherputs forward a new fixed-complexity SD based on soft tree pruning method.It uses reliabilityinformation on bits to estimate a search radius and thereby prunes unpromising nodes in the earlystage of the tree search detection to reduce complexity. Based on the soft tree pruning method, thefixed-complexity sphere decoder can still work in a fully pipelined mode. Simulation results showthat the soft tree pruning based fixed-complexity sphere decoder can provide a better trade-offbetween complexity and performance in interference-limited scenarios.