WSEAS Transactions on Systems
Print ISSN: 1109-2777, E-ISSN: 2224-2678
Volume 17, 2018
A New Time Adjusting Step-Size LMS Technique for Noise Cancellation Framework with Mean Square Deviation Analysis
Authors: ,
Abstract: In order to mitigate the negative impact of a non-stationary signal on the adaptive filter coefficients and to improve the trading off between convergence rate and steady-state error at various input levels, a new time adjusting step size LMS algorithm is proposed. The theoretical analysis of its mean square deviation (MSD) is investigated in this paper. Its closed-form expressions of mean square deviation for the transient and steady-state stages are also estimated. This new approach aims to reduce the steady-state MSD of the coefficients at high levels of input power signal without a scarifying on the speed of convergence as common in the conventional LMS and other VSSLMS approaches. We do so by developing an individual time adjusting step size based logarithmic function for each tap of the adaptive filter. It is found that the steady-state MSD depends directly on the minimum step size value when the reduction rate of step size is faster than the changing rate of the optimal coefficients. Based on the implementation of the adaptive noise cancellation, simulation results show the superiority of the proposed technique in term of possessing the lowest MSD at various input variances compared with others. Moreover, the proposed technique outperforms the compared algorithms in the matter of tracking a time-vary noise channel.
Search Articles
Keywords: Adaptive noise canceller, LMS algorithm, Variable step size LMS algorithms, Mean square Deviation MSD
Pages: 264-275
WSEAS Transactions on Systems, ISSN / E-ISSN: 1109-2777 / 2224-2678, Volume 17, 2018, Art. #30