Family of affine projection adaptive filters with selective partial updates and selective regressors

In this study the concepts of selective partial updates (SPU) and selective regressors (SR) in the affine projection adaptive filtering algorithm are combined and the family of affine projection algorithms (APAs) with SPU and SR features are established. These algorithms are computationally efficient. The mean-square performance of the presented algorithms are analysed based on the energy conservation arguments of Sayed's group. This analysis does not need to assume a Gaussian or white distribution for the regressors. The authors demonstrate the performance of the presented algorithms through simulations. The good agreement between theoretically predicted and actually observed performances is also demonstrated.