动态变化混叠模型下盲源分离中的源数估计  

Estimation of Sources Number for Underdetermined Blind Source Separation of Dynamic Mixing

作　　者：白琳[1] 温媛媛[1] 李栋[1] BAI Lin;WEN Yuanyuan;LI Dong(National Key Laboratory of Science and Technology on Space Microwave,China Academy of Space Technology(Xi’an),Xi’an 710100,China)

机构地区：[1]中国空间技术研究院西安分院空间微波通信全国重点实验室,西安710100

出　　处：《电讯技术》2024年第3期396-401,共6页Telecommunication Engineering

基　　金：稳定支持基金资助项目(HTKJ2021KL504005);空间微波技术重点实验室基金(6142411332103)。

摘　　要：在进行欠定盲分离时,特别是对于源信号数目及混合矩阵动态变化的情况,常规的欠定盲分离及源数估计方法不能对源信号数目的变化时刻做出判断,因此很难实现动态变化的源信号数目实时和准确的估计。针对这个问题,提出了一种动态变化混叠模型下欠定盲源分离中的源数估计方法。首先,建立动态变化混叠情形下盲源分离的数学模型及动态标识矩阵。其次,基于构建的动态标识矩阵统计和判断动态源信号数目的变化情况。最后,通过分段时间内多维观测矢量采样点聚类区间局部峰值统计,实现动态变化混叠模型下盲源分离中的源信号数目的有效估计。仿真结果表明,该方法能有效实现动态变化混叠模型下欠定盲源分离中的源数估计,并且信号估计效果良好。For dynamic sources and mixing matrix,conventional approach of blind signal separation and estimation of source number is difficult to accomplish the estimation of sources number for underdetermined blind separation because the change time of the sources cannot be estimated.To solve the problem,a method of sources number estimation for underdetermined blind separation of dynamic mixing is proposed.Firstly,the mathematical model of underdetermined blind separation of dynamic mixing and the dynamic identity matrix is established.Secondly,the change of the number of dynamic source signals is counted and judged based on the constructed dynamic identity matrix.Finally,the estimation of sources number of underdetermined blind separation of dynamic mixing sources is effectively accomplished based on the local peak statistics of multidimensional observation signals sampling data behind clustering at every time slice.Simulation results show that the method can effectively realize the estimation of sources number of underdetermined blind separation of dynamic mixing sources,and it has good performance.

关 键 词：欠定盲源分离 源数估计 标识矩阵 

分 类 号：TN911.7[电子电信—通信与信息系统]