空间调制信号的改进M-ML检测算法

空间调制(SM)系统的最大似然(ML)最优检测算法的计算复杂度很高,具有较低计算复杂度的M-ML检测算法受到了人们的关注.M-ML算法按照接收天线序号由小到大的顺序进行检测,从误比特率性能角度考虑并不是最佳的.通过研究不同检测顺序对算法性能的影响,提出了两个改进的M-ML算法,仿真结果表明改进的M-ML算法在误比特率性能上优于M-ML算法.由于M-ML算法在不同的信噪比下每层保留固定的节点数M,尤其在高信噪比时会造成计算资源的浪费,因此提出一种动态M-ML算法,即通过门限值自适应选择每层保留的节点数.仿真结果表明动态M-ML算法降低了M-ML算法的计算复杂度,同时性能逼近M-ML算法.

Improved M-ML algorithms for spatial modulation signal detection

In spatial modulation (SM) system, the computational complexity of the maximum likelihood (ML) optimal detection algorithm is very high. The M-algorithm to maximum likelihood (M-ML) detector has attracted increasing attention due to its lower computational complexity. However, in M-ML algorithm, the transmitted signals are detected according to ascending order of received antenna index, which is not the best way in the view of bit error rate (BER) performance. By studying the impacts of the different detection orders on the BER performance, two improved M-ML algorithms are proposed. The simulation results show that the proposed improved algorithms have better BER performance than M-ML algorithm. Moreover, in the M-ML algorithm, the number of retained nodes, M , is the same under different signal-to-noise ratio (SNR), which is unnecessary, especially at high SNRs. Thus, a dynamic M-ML algorithm is proposed which can adaptively change M by threshold to reduce the computational complexity while the BER performance is almost the same to M-ML algorithm. The simulation results also verify the advantages.