采用过采样技术的Infomax盲源分离算法

摘要

针对混合平稳信号的盲分离问题,提出了一种基于过采样技术的新盲源分离算法.对接收混合信号进行过采样,使接收平稳信号具有循环平稳特性.根据信息最大化算法,以输出信号的熵作为目标函数,将信号的循环相关函数和循环频率应用到分离矩阵的寻优中,实现信号的盲分离.仿真结果表明,该算法比传统的Infomax盲分离算法收敛速度快,收敛精度高.%Considering of the mixed stationary signal blind separation problem, a new blind source separation algorithm of cyclostationarity is proposed. The received signal and extracting its cyclostationarity property, take the output entropy as the objective function according to the Infomax theory, this new algorithm applied circulation related functions and circulating frequency into the separation matrix optimization to achieve the optimum blind separation. Compared the new algorithm with the traditional Infomax separation algorithm, the simulation results shown that the new algorithm convergent faster and has better convergence precision.