High-speed fir filter design and optimization using artificial intelligence techniques

Abstract

In this thesis, the design and optimization of high-speed FIR filters in signed powers-of-two (SPoT) space are discussed. Several non-linear optimization algorithms are developed based on artificial intelligence techniques. First, a systematic method based on an adaptive genetic algorithm is proposed for the design of high-speed FIR filters. The high-speed feature is achieved by factorizing a long filter into several cascaded subfilters each with SPoT coefficients. Significant reduction of hardware cost can be achieved. Second, two efficient algorithms are tailor-made for the design of very sharp linear phase FIR digital filters based on frequency response masking technique. Third, a hybrid algorithm formed by integrating genetic algorithm, simulated annealing and tabu search is proposed for the global optimization of FIR filters, which achieves not only the improved solution quality but also the considerable reduction of computational effort. Fourth, a modified micro-genetic algorithm is proposed to speed up the design process by utilizing a very small population pool.