A Variable Step Size Normalized Least-Mean-Square Algorithm Based on Data Reuse

The principal issue in acoustic echo cancellation (AEC) is to estimate the impulse response between loudspeaker and microphone of a hands-free communication device. This application can be addressed as system identification problem, which solved by using an adaptive filter. most common one for AEC normalized least-mean-square (NLMS) algorithm. It known that overall performance this algorithm controlled value its step size parameter. In order obtain proper compromise main criteria (e.g., convergence rate/tracking versus accuracy/robustness), specific term NLMS further designed variable represents motivation behind development algorithms. paper, we propose (VSS-NLMS) exploits data reuse mechanism, aims improve reusing same set (i.e., input reference signals) several times. Nevertheless, involved equivalent version NLMS, provides modes Based on approach, sequence sizes priori scheduled, advantageous terms computational complexity. simulation results context supported good features proposed VSS-NLMS