A new modified variable step size for the LMS algorithm

Adaptive FIR digital filters which use the LMS algorithm have found application in quite a wide area of digital signal processing due to their simplicity. In all these areas of application, the measurement of good performance has always been a faster convergence compounded with minimum misadjustment. However, with the LMS algorithm, a higher convergence speed is always complemented by a higher misadjustment. In trying to overcome this trade off, different kinds of variable step size algorithms and the modified variable step size algorithm have been proposed. Unfortunately these algorithms are affected by colored disturbance signal. In this paper we propose a new modified variable step size algorithm, which is independent of both white and colored disturbance noise