一种低复杂度的近似最大似然MI~/IO检测算法

Complexity reduction ML detection algorithm for MIMO system

摘要:

为保证无线传感器网络协作式V-BLAST传输中，在较高的检测性能的前提下大大降低算法复杂 度，提出一种低复杂度的近似最大似然检测算法．将传统的V-BLAST算法性能最好一层解的邻域作为候选 判决集合，并以此邻域内每一个符号作为初始值进一步采用传统的V-BLAST算法反馈判决其他层的符号， 采用最大似然准则对候选向量进行判断．该方法有效减小了最大似然检测算法检测向量的数目，降低了算法 的复杂度．仿真结果表明该算法具有良好的综合性能

Abstract:

Aiming at reducing the computational complexity greatly and achieving high detection performance in cooperative MIMO-based WSN, a new complexity reduction ML detection algorithm is proposed. Using con- ventional V-BLAST algorithm, the best performance layer is found, and the neighborhood is considered to be candidate set. Regard the every symbol in the candidate set as initial value, we adopt V-BLAST algorithm a- gain to detect the symbol of other layers. At last, we use the maximum likelihood criterion to judge the candi- date vector. Because the number of constellation point in the maximum likelihood detection algorithm is re- duced effectively, the complexity of the algorithm decrease greatly. The simulation results show that the pro- posed scheme obtains good comprehensive performance