Design of Low Power, High Performance FIR Filter Using Modified Differential Evolution Algorithm

摘要

In digital filters, maximum power consumption occurs during multiplication operations. Hence, to reduce power consumption, the number of multiplications has to be minimized. The number of Signed-Power-of-Two (SPT) terms in the filter coefficients has to be optimally minimized, without compromising on the filter response. A modified Differential Evolution (MDE) algorithm has been used to generate optimized coefficients for digital FIR filters. The filters designed using MDE and the standard Remez (REM) Exchange method are synthesized in 90nm technology and their performances are compared. The critical delay, area and power of MDE implementation are found to improve by 6%, 12% and 16%respectively with respect to REM.