A Novel Neural Networks-Based Approach for Designing FIR Filters

A novel linear-phase finite-impulse response (FIR) digital filters design approach based on neural networks optimization technique was proposed in this paper. The main idea is to minimize the weighted square-error function in the frequency-domain. The convergence theorem of the neural networks algorithm was presented and proved to illustrate the proposed algorithm stable, and the implementation of the approach was also described together with some design guidelines. This method can control the overshoot phenomenon that may happen near the pass-band and stop-band edge of the designed filter. Some optimal design examples were given to illustrate the effectiveness of the proposed optimal design method