The NP-VSS NLMS Algorithm with Noise Power Estimation Methods For Acoustic Echo Cancellation

The Variable Step-Size Normalized Least Mean Square (VSS-NLMS) algorithm types are currently a better technique in the field of adaptive filtering. These algorithms are based on the control of the adaptation step in order to solve the trade-off caused by the fixed step-size of the conventional NLMS algorithm. The VSS-NLMS algorithms have demonstrated a good performance of the filtering error in the steady state regime, by keeping the convergence speed and the tracking capability similar to the original NLMS algorithm. The non-parametric VSS-NLMS is one of the most used algorithm in this family, where the variable step-size is updated using a theoretical background noise power, the challenge is how to estimate this quantity which is unknown in practice. In this paper, we propose a new efficient method to estimate the noise power and then evaluate it with several other methods already proposed. After several modifications, simulation results show that all these methods are similar to the original NPVSS-NLMS in the stationary and the non-stationary conditions.