基于Gibbs采样的MIMO信号检测改进算法

针对传统Gibbs采样算法的“失控问题”,提出了一种改进算法，在Gibbs迭代采样过程中加入对噪声方差的估计,消除了失控问题。设计了信号检测流程,将最大似然（ML）检测的搜索过程嵌入Gibbs迭代采样过程,减少了最大似然检测的计算步骤。仿真结果表明,所提算法能估计噪声方差,具有接近最大似然检测的性能。

An improved MIMO detection algorithm based on Gibbs sampler.

For preventing the stalling problem of Gibbs sampler,an improved algorithm is proposed. Firstly,the estimates of the noise variance is obtained from the Gibbs sampler,the stalling problem is eliminated. Then,by combining with Maximum Likelihood(ML) detection in the process of iteration of the Gibbs sampler,the complexity of signal detection is reduced. Simulation results show that the noise variance can be obtained with the proposed method,and the improved algorithm achieves the performance approaching the maximum likelihood detection.