Research Of Adaptive Active Noise Control Based On Neural Network

With development of modern industry and transportation industry, the subject of noise control has attracted people's growing attention and concern. Adaptive active noise cancellation based on the adaptive control theory has become one of important researches in the field of active noise control and has gain a lot of achievement. The adaptive control algorithm which can directly affect the adaptive control quality is playing important role in adaptive active noise control system. At present, The Filter-XLMS algorithm which base on linear adaptive filter has been widely used in the field of active noise control. But, this algorithm has some shortcomings, which contain the linear adaptive filter should be a high order filter and the algorithm cannot deal with the nonlinear noise. Therefore, it is a very potential and meaningful work to improve the performance of control algorithm by means of new information technique.Focusing on those problems, This paper has given the model of feed-forward adaptive active noise control system. Adaptive active controller of this model is the BP neural network which replaced the linear adaptive filter. And I has proposed a new adaptive active noise control algorithm—improved BP Filter-x algorithm. This algorithm can easily deal with the nonlinear noise. Because of BP neural network existing the shortages of slow convergence speed and easily falling into local minimal value. Which cases improved BP Filter-x algorithm need a lot of computational time and has slow convergence speed.Genetic algorithm is a high-performance random search and optimization method. which has very good global optimization ability, can effectively jump out of local minimum points, and has good adaptability and high degree of parallelism. So I also a new adaptive active noise control algorithm—Genetic-BP algorithm, Which use the global optimization of genetic algorithm to optimize neural network weights, Which is approaching the globally optimal value. Then, we can get the globally optimal value with help of improved BP neural network. We can use the globally optimal value to control the adaptive active noise control system. And we also made a very good control effect.In order to verify the reducing noise effect of the algorithm, which has been given in this paper. We carried on the system simulation experiment, using filter-XLMS algorithm, improved BP filter-x algorithm, Genetic-BP algorithm separately, under linear or non-linear conditions. The simulation results are consistent with theoretical analysis.