Search

The conflict of filtered-x least mean square (FxLMS) algorithm between convergence speed and steady-state misalignment restricts the performance of the adaptive system. The convex combination approach can successfully avoid this contradiction as it combines two adaptive filters that adopt different step sizes. Firstly, the generalized versoria function is applied as the cost function to develop the filtered-x maximum versoria criterion (FxMVC) algorithm, which has good performance in the face of impulsive noise and performs better than the FxLMS algorithm. Then, inspired by the convex combination approach, C-FxMVC is developed by combining the FxLMS algorithm with the FxMVC algorithm to obtain better behavior, and the FxLMS algorithm is employed in the fast filter. To further optimize the performance of the presented algorithm, we use the method of amplitude-constraint, and the amplitude of e (n) and X (n) are restrained nicely. Numerous simulations are employed to verify the effectiveness and rationality of the presented algorithm, where the impulsive noise required for the simulation environment is consistent with symmetric α -stable (S α S) distribution model. Simulation results demonstrate that the proposed C-FxMVC algorithm can possess better efficacy in terms of convergence speed and noise reduction. [ABSTRACT FROM AUTHOR]