幂函数的稳定性和有效性在盲分离算法中的应用

The stability and efficiency analysis of power function in the BSS algorithm

涉及盲分离问题的各种算法在实现的过程中,针对不同的源信号选取合适的幂函数参数直接影响到分离结果的优劣,本文从两方面对幂函数参数对算法的影响进行讨论.首先分析幂函数和算法稳定性的关系.找出幂函数针对不同源信号的相应参数取值;然后分析幂函数和算法有效性的关系,给出幂函数参数变化对分离结果的影响。然后将幂函数的稳定性和有效性具体化地应用到盲分离迭代算法中,扩大了分离算法对源信号的适用性,可以提高信号分离的准确性以及算法本身的稳定性。

ln the implement process of BSS algorithm, how to choose the precise power function for the different source signals will affect separating result directly. This paper will discuss the affection of the power function in two aspects. First, the relationship between the power function and the algorithm stability will be analyzed, finding the corresponding parameter values of the power function for the different source signals, and then the relationship between the power function and the algorithm efficiency will be analyzed, showing the affection on the separating result when the parameter of the power function changing. Then the stability and the efficiency are used to the BSS iterative algorithm. The algorithm applicability is extended, the veracity of signal separation and the stability of BSS algorithm is improved.