The design of low-cost fixed-point FIR digital filters using an improved local search

This paper presents a method for the design of low-cost field programmable gate array (FPGA) based finite duration impulse response (FIR) digital filters. A discrete Lagrangian local search is used to minimize the hardware cost (the total number of required FPGA logic elements) subject to constraints on the passband to stopband energy ratio (PSR) and on the maximum passband and stopband ripple. Previous discrete Lagrangian local search methods do not produce minimum hardware cost designs and may not even converge on a solution for this design problem.