Simple variable step size LMS algorithm for adaptive identification of IIR filtering system

This paper proposes a simple variable step size Least Mean Square (LMS) algorithm for adaptive identification of Infinite-Impulse-Response (IIR) filtering system. The proposed algorithm is called Fast Variable Step Size LMS (FVSSLMS) which incorporates a recursively variable adaptation step size based on error square multiplying by a constant. The simulation results show better performance than traditional LMS and Normalized LMS (NLMS) algorithm in terms of fast convergence time and less misadjustment in a steady state.