Grading learning for blind source separation
摘要
By generalizing the learning rate parameter to a learning rate matrix, this paper proposes agrading learning algorithm for blind source separation. The whole learning process is divided into threestages: initial stage, capturing stage and tracking stage. In different stages, different learning rates areused for each output component, which is determined by its dependency on other output components. Itis shown that the grading learning algorithm is equivariant and can keep the separating matrix from be-coming singular. Simulations show that the proposed algorithm can achieve faster convergence, bettersteady-state performance and higher numerical robustness, as compared with the existing algorithmsusing fixed, time-descending and adaptive learning rates.