Step-Size Optimization EASI Algorithm for Blind Source Separation

Abstract

Aiming at the problem of blind source separation of the communication signals, we propose a step size optimization equivariant adaptive source separation via independence (SO-EASI) algorithm basing on the EASI block based algorithm. This algorithm adjusts the step-size by the steepest descent method and thereby greatly increases its convergence speed whatever value the step-size is initialized. Simulation results show that SO-EASI algorithm can effectively blindly separate the communication signals and these results also support the expected improvement in convergence speed of the approach.