Speech enhancement using convex combination of fractional least-mean-squares algorithm

In this paper, we propose a novel method, called convex combination of fractional LMS (CC-FLMS) for dual-channel speech enhancement based on a modified version of LMS algorithm. Here, we use a modified Least-Mean-Squares algorithm, known as FLMS, by incorporating the fractional derivative term in the weight adaptation equation of the standard Least-Mean-Squares (LMS) algorithm. Combinational approaches provide an interesting way to improve adaptive filtering performance. We use the convex combination of two transversal filters for speech enhancement. The individual filters, designed by the FLMS algorithm, are independently adapted using their own error signals, while the combination is adapted by means of a stochastic gradient algorithm to minimize the error of the overall structure. Experimental results show that the proposed method has a better performance in dual-channel speech enhancement in a sense of mean-squares error (MSE) and speech quality improvement than the LMS, FLMS, and CC-LMS algorithms.