Digital filter design using LMI based techniques.

Abstract

Linear Matrix Inequalities (LMI) basd digital filter design algorithms are developed and applied to the FIR (one dimensional) optimal filter design problem and FIR (one dimensional) power spectrum estimation problem. The IIR (one dimensional) LMI based digital filter design and IIR (one dimensional) power spectrum estimation problems are formulated as linear objective optimization problems with LMI constraints and solved using an alternative scheme to overcome the non-linearity optimization formulation. In addition, the LMI based formulation of the filter design problem is applied to general approximation/model reduction problems (SISO). The solutions obtained by solving several examples show better results as compared with the traditional design methods.