论文部分内容阅读
为解决卷积混合盲信号分离时域方法收敛速度慢的问题,提出了卷积情况下的非线性主分量分析(PCA)准则,并分析其与高阶统计量准则之间的等价关系,推导了一种解决卷积盲信号分离问题的非线性PCA方法。作为一种递推最小二乘(RLS)类型的算法,所提方法与现有的自然梯度算法和高阶统计量算法相比具有收敛速度快、跟踪性能好的优点,计算机仿真实验验证了算法的有效性。
In order to solve the problem of slow convergence rate of convolutional mixed blind signal separation time domain, a nonlinear principal component analysis (PCA) criterion in convolution case is proposed and its equivalence relation with higher order statistics criterion is analyzed. A nonlinear PCA method for solving the problem of separation of blind signals is derived. As a Recursive Least Squares (RLS) type algorithm, the proposed method has the advantages of fast convergence rate and good tracking performance compared with the existing natural gradient algorithm and high-order statistics algorithm. The computer simulation results show that the algorithm Effectiveness.