Design of sparse FIR filters using linear programming

More efficient use of multipliers in finite impulse response (FIR) filters can be achieved at the expense of a slight increase in delay by designing sparse FIR filters. Two new methods of designing sparse linear phase FIR filters are reported. One results in sparse direct-form filters for a particular class of filters, and is based on Fourier analysis. The second method is heuristic in nature, and gives a relatively simple approach to designing sparse cascaded filters. Both methods use linear programming for the design