【摘要】 针对粒子群对盲信号分离时出现早熟等现象,提出负熵粒子群算法.首先将信息负熵最大化作为粒子群的目标函数,不依赖其它的非高斯性度量判断,避免识别混叠矩阵;然后对观测信号进行中心化和白化处理,用分离矩阵调整,使各个信号分量之间独立,大权重粒子做全局搜索,小权重的做局部搜索,混叠矩阵对所有列元素数组完成分离;最后给出了算法流程.MATLAB仿真结果显示该算法能够有效地完成盲信号分离和主要参数的提取.更多还原

【Abstract】 Aiming at the precocity in Particle Swarm Optimization(PSO) Algorithm,proposes the negative entropy PSO Algorithm.First the information negative entropy maximization as the objective function of PSO,without the other non-Gauss measure judgment,avoids distinguishing the aliasing matrix;Then carries on the centralization and albinism processing to the observation signal,causes between each signal component with the separation matrix adjustment the independence,the great weight makes the overall situation search,the small weight makes the partial search,the aliasing matrix completes the separation to all row element array;Finally it given the Algorithm.The MATLAB simulation result showed this Algorithm can complete the blind signal separation and the main parameter extraction effectively.更多还原