Extrapolated LMS adaptive filtering algorithms

Two novel adaptive filtering algorithms are proposed in this work. The algorithms greatly reduce the number of adaptation steps of the LMS algorithm by extrapolating the adaptive filtering coefficients. The extrapolation is done through the time constant and optimal coefficient estimations of the coefficient adaptation process. Simulation results for the system identification and correlation canceller loop show significant reduction of the adaptation steps. The algorithms can be applied to general orthogonal adaptive filters and least-square error optimization problems.