%A YANG Hua, ZHANG Hang, ZHANG Jiang, YANG Liu and LI Jiong %T Initial Separating-matrix Optimized Online Blind Source Separation Algorithm %0 Journal Article %D 2016 %J Computer Science %R 10.11896/j.issn.1002-137X.2016.6A.063 %P 265-267 %V 43 %N Z6 %U {https://www.jsjkx.com/CN/abstract/article_3359.shtml} %8 2018-11-14 %X Aiming at the problem that the convergence speed of online BSS algorithms is effected by the initial separating matrix,a new initial separating matrix optimized online BSS algorithm based on ABC algorithm was proposed.The new algorithm utilizes the strong search ability of batching ABC algorithm to get an optimized initial separating matrix that makes the BSS algorithms have a better initial iteration point,and then uses the gradient descent to achieve online separation,which can improve the overall convergence speed of BSS algorithm.Simulation results confirm that the proposed algorithm increases the convergence speed effectively.And in the condition of time-varying mixing matrix the new algorithm is applicable.