A novel method for overdetemined blind source separation
Yunfeng Xue, Yuja Wang, Qiudong Sun · 2010
In this paper, the blind separation of overdetermined mixtures is considered, i.e., the case where more sensors than sources are available. The contrast function for overdetermined blind source separation problem is studied, and the gradient of the contrast function is presented. Based on the gradient, a novel learning algorithm is proposed to solve the overdetermined blind source separation problem. A computer simulation example demonstrates the performance of the proposed algorithm. Index Terms—Blind source separation; Independent compo- nent analysis; Contrast function