Application Study of Neural-Network Parallel Algorithm in the Optimal Design on FIR Multi-Band-Stop Filters

摘要

This paper traverses the relations between the amplitude-frequency response of FIR filters with the linear phase and the parallel algorithm of neural network, which its activation matrix is H_d — C~ T W , presents and proves the convergence theorem of the neural-network parallel algorithm, and gives the optimal design examples of high order FIR multi-band-stop digital filters. The results show that the parallel algorithm is efficient in the field of optimal design on the high order FIR multi-band-stop digital filters with the linear phase.