Adaptive filter realization with a minimum number of multipliers

A new approach to the realization of adaptive FIR filters with a minimum number of multipliers is presented. The realization requires only 3 multipliers for any filter length, incorporating a full coefficient updating scheme at every iteration which allows for fast convergence. The adaptation uses the least mean square (LMS) algorithm with a modification in order to reduce the number of multipliers of the filter.