Recursive digital filter design in the complex domain using an efficient method

Presents a generalized Ellacott-Williams (GEW) algorithm for the design of recursive digital filters in the complex domain. Two design procedures which are two modified versions of the original Ellacott-Williams algorithm are first developed. At each iteration, linear complex Chebyshev approximation problems are formulated based on the least square error criterion to find the solutions for the numerator and denominator. The approximation problems can be solved by using a weighted least squares (WLS) algorithm proposed by Y.C. Lim et al. (1992). This provides an efficient design technique since the heavy computational burden of using a linear programming algorithm can be solved. The main differences between these two design procedures are outlined. Computer simulation results show that the proposed GEW algorithm has better design capabilities than other methods