Research On Active Noise Countermeasure Control Algorithms

In traditional active noise control techniques,the noise to be cancelled is often assumed to be a single signal,but in many cases,useful signals may exist in the anechoic region,and if active noise control techniques are used blindly,the useful signals will be cancelled at the same time in the noise reduction process,affecting the system performance.In this paper,through the study of the traditional active noise control system,according to the theory of active noise control technology,an active noise countermeasure control system with the function of canceling the interference signal,extracting the useful signal and improving the signal-to-noise ratio is proposed,based on which the noise active control algorithms of linear correlated noise and non-linear correlated noise are studied respectively.First,this paper addresses linear correlated noise.In order to solve the contradiction between the convergence speed and steady-state error of the traditional Filtered-X Least Mean Square(FXLMS)algorithm,an improved variable step size FXLMS algorithm based on the Softsign function is proposed,and the simulation results show that the improved algorithm has a faster convergence speed and smaller steady-state error than the existing variable step size algorithm.In order to further reduce the steady-state error of the improved variable step size FXLMS algorithm,this paper proposes an improved convex combined FXLMS algorithm,which combines the improved variable step size FXLMS fast filtering algorithm based on the Softsign function with the FXLMS slow filtering algorithm with a small step size factor.The simulation results show that the proposed improved convex combined FXLMS algorithm outperforms the single filtering algorithm in terms of convergence speed and steady-state error,and has a better noise reduction effect in the active noise countermeasure control system.Then,for non-linear correlated noise,on the basis of Error Back Propagtion(BP)neural network,this paper builds a BP neural network model based on improved particle swarm algorithm and further proposes an improved neural network active noise countermeasure control algorithm.By adjusting the parameters of inertia weights and learning factors,the proposed improved particle swarm optimization algorithm has a better optimization finding ability and convergence accuracy.The simulation results show that the proposed algorithm outperforms the traditional BP neural network and the traditional particle swarm optimised BP neural network active noise countermeasure control algorithm in terms of noise reduction of non-linear correlated noise.Finally,in order to simplify the simulation operation process of the active noise countermeasure control algorithm and further test the noise reduction performance of the improved algorithm,this paper uses MATLAB to design an active noise countermeasure control system simulation platform,which combines the improved variable step size active noise countermeasure control algorithm and the improved neural network active noise countermeasure control algorithm,and has the ability to handle both linearly correlated and nonlinearly correlated noise,and the main functions of the simulation platform include signal reading and writing,waveform display,noise reduction processing and calculation of signal-to-noise ratio.