摘要
文章针对基于因子图置信传播(BP)算法应用于大规模MIMO检测中存在复杂度高以及环路问题,提出了基于高斯树近似的置信传播(GTA-BP)算法。通过高斯近似在消息传递过程中降低算法的复杂度。对于循环因子图,利用Prim算法找到最优近似树,将BP算法应用到最优近似树上,能解决因环路导致算法不收敛问题。仿真结果表明:GTA-BP算法在降低算法复杂度的同时能避免因循环而无法准确收敛的情况。
Aiming at the high complexity and loop problem in the large-scale MIMO detection based on the factor graph BP algorithm,GTA-BP is proposed in this paper. The complexity of the algorithm is reduced during the message passing process by Gaussian approximation. For loopy factor graph, Prim algorithm is used to find the optimal approximation tree, and the BP algorithm is applied to the optimal approximation tree to solve the problem that the algorithm does not converge because of the loop. The simulation results show that GTA-BP algorithm can reduce the complexity of the algorithm while avoiding the inability to accurately converge due to the loop.
引文
[1]PARK S,CHOI S.QR decomposition aided belief propagation detector for MIMO systems[J].Electronics Letters,2015(11):873-874.
[2]TANABE S,SHIGYO A D,ISHIBASHI K.Not-so-large MIMO signal detection based on damped QR-decomposed belief propagation[C].Monterey:International Symposium on Information Theory and ITS Applications,IEEE,2016:463-467.
[3]ZHANG Y,GE L,YOU X,et al.Belief propagation detection based on max-sum algorithm for massive MIMO systems[C].Nanjing:International Conference on Wireless Communications&Signal Processing,IEEE,2017:1-6.
[4]GAO Y,NIU H,KAISER T.Massive MIMO detection based on belief propagation in spatially correlated channels[C].Hamburg:The11th International ITG Conference on Systems,Communications and Coding,2017:1-6.