Weighted least squares design of FIR filters subject to magnitude and phase constraints

This paper proposes an algorithm for the design of FIR digital filters with arbitrary magnitude and phase responses. The designed filters are optimum according to a weighted least squared frequency domain error criterion with constraints on the resulting magnitude and phase responses. This optimality criterion offers more flexibility than pure least squares or Chebyshev criteria. These widely used criteria are included as special cases. The proposed algorithm computes the optimum solution of the design problem by solving a sequence of quadratic programming problems. It is efficient and numerically stable.