An affine projection algorithm with decomposition for improved performance

An Affine Projection algorithm (APA) with decomposed weight vector and variable step size is proposed in this paper so as to improve the performance of adaptive filter. The optimal value of variable step size is obtained by the largest decrease in the mean square deviation error. This proposed variable step size optimally decomposed APA has improved convergence with smaller value of steady state error. Moreover the total computational complexity, time and cost is also found to be less. The proposed algorithm is validated by simulation for acoustic echo cancellation application. From the simulation results, it is proved that the proposed algorithm has faster convergence and smaller steady state error than conventional APA.