Design of Arbitrary Complex Coefficient WLS FIR Filters With Group Delay Constraints

This correspondence presents a new method for the weighted least-squares (WLS) design of complex coefficient finite-impulse response (FIR) digital filters with arbitrary frequency responses by imposing an explicit group delay constraint. The nonlinearity of the group delay with respect to the filter coefficients is circumvented by linearization so that the group delay constraint can be successfully incorporated into the second-order cone programming (SOCP) framework. The effectiveness of the proposed method is illustrated by two nontrivial examples and compared favorably with some existing methods.