Application Of BP Neural Network Based On Genetic Algorithm And LM Algorithm To The Design Of Digital Filter

BP neural network, as an artificial intelligence technique, develops rapidly in the recent fifty years, and attains compelling production. In the recent few years, many researchers at home and abroad have applied it to design digital filter, and utilize the function approach capacity of BP neural network to approach the ideal system function of the filter, and attain excellent effect. But during the using of BP neural network in designing filter, we find there are many localizations, such as the convergence to local least value,unable convergence to the given error; the studying period is long,the time of iteration is more and the speed of convergence is slow; the choice of the first connecting value is blind and without overall situation and . Those shortages will depress the effect of the network studying, so we must adopt some corresponding metheods to improve the simulating effect and forecasting precision.Point to the above problems of BP neural network, in this paper, BP neural network based on Genetic algorithm and Levenberg-Marquardt algorithm is introduced. Levenberg-Marquardt algorithm is a rapid algorithm using the standard numeric optimizing technique. Appling Levenberg-Marquardt algorithm to the BP neural network can solve the convergence to lacal least value and accelerate the speed of convergence. Genetic algorithm is an arithmetic based on evolution and genetics used to search the optimization. It is simple and robust. Using Genetic algorithm to optimize connecting value can increase the efficiency of the function approach capacity of BP neural network.Then BP neural network based on Genetic algorithm and Levenberg-Marquardt algorithm is used in designing FIR digital filter. After comparing with only BP neural network, filter designed by the new algorithm has the less passed band attenuation and the more stopped band attenuation, and reach an excellent optimizing designing result.