An efficient MIMO detection algorithm with the resistance to channel correlation

MIMO,the core technology of next generation wireless broadband communication,is faced with the contradiction between the system gain benefit by increasing antennas and the detection error due to high channel correlation.Therefore,we propose a new Lattice-Reduction(LR)-assisted K-Best algorithm.Since the children nodes of every parent node are uncertain after LR processing,on-demand expansion scheme is adopted to expand children nodes,and the sorting algorithm based on minimum candidate heap is used to decrease the sorting complexity.Then the average time complexity is decreased from O(KNlog2(KN)) to O(Klog2K),and the spatial complexity is decreased from O(KN) to O(K).In order to solve the problem of detection error caused by no longer limited constellation map after LR processing,a boundary control scheme is put forward to improve the detection accuracy.Simulations show that with the aid of boundary control,the BER performance has obtained 3dB gain when channel correlation is high.And the proposed algorithm only lost 1dB gain compared to ML algorithm with much smaller computational complexity,the complexity is only linearly increased with the number of antennas.So this proposed efficient algorithm is suitable for large scale antenna system.