源信号空间分布对欠定盲源分离可辨识性影响的分析

摘要

This paper mainly studies the effect of source location on identifiability of underdetermined blind source separation.Based on the identifiability of linear system,the paper models the problem of underdetermined blind source separation using a linear array.Theoretical analysis shows that an underdetermined system cannot be identified when sources are located in some areas and the problem of underdetermined blind source separation cannot be resolved.The areas where an underdetermined system cannot be identified are reconfirmed after the cramér-rao bound for estimating the mixing matrix in noisy sparse component analysis is deduced.Simulations confirm the theoretical analysis.%本文主要研究了欠定盲源分离中源信号的空间分布对系统可辨识性的影响.从系统可辨识性问题出发,对瞬时欠定盲源分离问题进行空间建模,理论分析表明,当源信号在空间处在某些特定区域时,系统是不可辨识的,从而无法正确实现欠定盲源分离.利用混合矩阵估计的克拉美罗界,进一步明确了影响系统可辨识性的源信号空间分布区域.仿真实验验证了理论分析的正确性.