The Application and Simulation of Block Varying Step Size FDLMS Adaptive Algorithm for System Identification Problem

To solve the problem that the convergence performance of classical adaptive filtering algorithms is sensitive to input signal power and is hard to balance between convergence speed and steady-state misadjustment, this paper, based on the traditional Frequency-Domain Block Least Mean Square (FDLMS) algorithm, presents a variable step size which is easy and reliable tuning parameter is controlled by current input signal energy and filtering error energy together. We propose a bin-wise block-varying step-size for the FD least-mean-square (LMS) algorithm. It achieves both fast convergence rate and low steady state error. In addition, simulation results for adaptive filtering are presented to demonstrate the performance improvements in convergence speed and steady-state misadjustment compared with other existing frequency domain algorithms such as the Gradient constrained frequency domain adaptive filter and unconstrained frequency domain adaptive filter for real valued data.