A novel adaptive algorithm has been proposed that provides faster convergence than the least mean square (LMS) and normalized LMS (NLMS) algorithms. In the proposed algorithm, the error sequence is used for designing the step size parameter at each iteration. A strategy to reduce the computational complexity required in the NLMS adaptation is also used. Convergence analysis of the proposed algorithm is made in a comparative fashion to the NLMS algorithm. Computer simulations demonstrate that in the scenario of channel equalization, the proposed algorithm based on the structure of a transversal filter accomplishes faster start-up than the combined LMS/F algorithm recently addressed.
展开▼