Efficient weighted L/sub P/ algorithm for the design of low-pass FIR filters with some closed-form derivations

An efficient iterative reweighted least-squares (IRLS) algorithm for the design of L/sub P/ approximation FIR filters is investigated. The optimal filter coefficients are obtained by iteratively solving a system of linear equations whose associated matrix can be further simplified as closed-form expressions and pre-calculated as the filter is specified. Simulation results indicate the excellent performance of the proposed algorithm has a complexity of O(N/sup 2/).