Digital filters with peak-constrained weighted least-squared errors

Digital filter design algorithms that extend peak-constrained weighted least-squared error optimization methods to cover a wide range of problems are presented. It is shown that the optimization problem can be solved in terms of the equivalent problem of minimizing a modified total weighted-square error function. Several different approaches to determining the modified weighting function using iterative reweighted least-squares methods are discussed. A generalized multiple exchange algorithm for designing FIR digital filters and windows is also discussed