Research And Application Of Multi-channel Active Noise Control Algorithm

Noise pollution is a common phenomenon in life.High-frequency harsh continuous noise can damage the auditory system.Continued low-frequency noise can affect the nervous system,leading to low attention and poor quality of life.The traditional way to reduce noise is to use sound-absorbing materials,but this method has little effect on low-frequency noise.Active noise control is a new noise reduction technology which is a combination of acoustics,adaptive filtering,digital signal processing and other disciplines.Compared with the traditional passive method,the active noise control system has become the main research direction of noise elimination with its good adaptability,low cost and diversity of applicable environment.The multi-channel active noise control algorithm is evolved from a single-channel algorithm for noise cancellation in a three-dimensional environment.Due to the complexity of the noise situation in the three-dimensional space,it is necessary to use the multi-channel microphone and the speaker to collect and cancel the stereo space noise signal.Resulting in a higher amount of computation,affecting the real-time performance of adaptive filters.Therefore,based on multi-channel scenario,this paper studies a variety of active noise control algorithms.This paper analyzes and studies the large computation and slow convergence speed of multi-channel algorithm.In order to improve the convergence rate of the algorithm,we analyzed the range of the convergence factor,steady-state and transient properties of the algorithm.Focus on least mean square algorithm(LMS),the influence of parameters on the performance of the algorithm is analyzed in detail.In order to improve the noise reduction of active noise control,we should consider the problem of signal delay encountered in the real environment,so filtered-x least mean square algorithm(FXLMS)is researched.To solve the acoustic delay caused by phase imbalance,we need model the secondary path.This paper proposed an improved FXLMS algorithm structure,which reduces the amount of computation.The secondary channel is identified offline,and the estimation parameters are applied to the improved FXLMS algorithm to verify the dual channel active noise control system.Based on the theory of LMS algorithm,the principle and performance of affine projection algorithm are studied and the corresponding improvement methods are proposed.Affine projection algorithm is a kind of signal reuse algorithm,which has a good convergence speed compared with other adaptive filtering algorithms.There are two factors that affect the performance of the algorithm,step size and projection length.In this paper,I propose a new variable step size affine projection algorithm(VSS-APA).It dynamically changes the step size according to certain rules,so that it can get smaller steady-state error and faster convergence speed.The simulation results can prove that the performance is superior to the traditional affine projection algorithm,and in the active noise control(ANC)application,this new algorithm can get very good results.