基于传感器阵列的近场时延信号盲源分离算法

Blind Source Separation for Near Field Slightly Delayed Mixtures Based on Sensor Array

主要研究了近场条件下,当接收阵列信号存在微小时延的盲分离问题。此情况下传统的线性无记忆盲分离模型已不再适用,提出了一种通过引入源信号的导数并改进混合模型后,利用二阶统计特性的算法,可以轻松解决时延条件下的盲分离问题(BSS)。同时当传感器数远多于源信号数时,该算法还能达到消噪的目的,仿真结果证明了算法的有效性。

The problem of blind source separation （BSS） of different time delays in each sensor in near field is studied. The linear memoryless BSS model is no more suitable in this case. When time delays are small in comparison to the coherence time of each source, the research shows that BSS can be simplified to extract the sources by building up a particular set of instantaneous mixtures involving derivatives of sources with respect to time and by using an adapted second order statistical property. When more sensors than sources are available, this algorithm can also be used to denoise. The computer simulationproves the validity of this algorithm.