基于核独立成分分析的盲多用户检测算法

针对部分多用户检测算法需要对信道参数进行估计的缺点，提出了一种基于核独立成分分析的盲多用户检测算法．该算法根据源信号的不同分布情况，在重建核希尔伯特空间内选取不同的非线性函数作为对比函数，将信号从低维空间映射到高维空间．在高维空间，接收端利用已知信息，将目标用户扩频码作为解混矩阵的初始值，利用自适应方法进行迭代，有效地解决了盲信号分离的无序性，实现了目标用户信号的提取．仿真实验表明，该算法的误码率性能在用户数量增大和远近效应严重的情况下都远优于基于匹配滤波器的单用户检测器，与传统独立成分分析方法相比更具灵活性和鲁棒性．

Blind Multiuser Detection Based on Kernel Independent Component Analysis

To avoid the drawback that channel parameters were assumed to be known in some existing methods, a new blind multiuser detection approach for the direct sequence code division multiple access system based on kernel independent component analysis (KICA) was presented. The desire user's symbols were effectively separated from the received signal by adopting nonlinear functions chosen adaptively from the reproducing kernel Hilbert space to transform the signal into the high dimension space, initiating the KICA demix matrix with the desire user's signature sequence, and iterating the demix matrix with adaptive methods. Simulation results show that, compared with the conventional match filter detector, the new detector provides better bit error rate performance, especially when the user number is large and the near-far problem is severe, and it is more flexible and robust than those using the conventional independent component analysis methods.