DATA-DEPENDENT PARTIAL UPDATE ADAPTIVE ALGORITHMS FOR LINEAR AND NONLINEAR SYSTEMS (TueAmOR12)

Author(s) : Tyseer Aboulnasr (University of Ottawa, Canada) Qiongfeng Pan (University of Ottawa, Canada)

Abstract : In this paper, we will review partial update adaptive algorithms with special emphasis on data-dependant algorithms. We then demonstrate that the same approach applied in the MMax LMS partial update algorithm for linear adaptive filters [4] can be extended to the class of nonlinear filters known as Volterra filters. The impact of the fact that the input vector is no longer a set of delayed input values on the complexity reduction due to the partial update is noted. Simulation results show that, as for linear filters, considerable saving is possible with little deterioration in performance.

Menu