Research On Active Control Algorithm For Impulsive Noise

Noise,which can cause people to feel irritable,has become one of the four major contemporary environmental pollutants,seriously jeopardising human physical and mental health.The traditional noise reduction techniques(sound absorption,sound isolation,sound elimination)show excellent noise reduction effects for high frequency noise,but are not as effective as they could be for low frequency noise.As a result,active noise control strategies are attracting more and more attention.Although active noise control techniques have made great progress in recent years,it is difficult to obtain good noise reduction effects in practical applications due to the complexity of the noise components.Among them,noise reduction of random noise with impulsive is difficult to achieve the desired effect.Impulsive noise is a form of noise that produces a large magnitude at a certain moment and has a short duration.Compared to continuous noise,impulsive noise is more harmful to the human body.Therefore,this paper explores and researches active control methods for impulsive noise,and the following works are accomplished:Firstly,this paper provides a brief introduction to the theoretical basis of active noise control.Secondly,the study summarizes the research and application status of active noise control algorithms,especially the research for impulsive noise control,such as Fx LMP algorithm and Fxatan LMS algorithm.In this paper,we also analyze the advantages,disadvantages and noise reduction performance of the existing impulsive noise control algorithms.On the basis of the above analysis,this paper proposes the C-Fx MVC algorithm based on an improved convex combination structure.The algorithm uses a combination of the Fx LMS algorithm and the Fx MVC algorithm,where the Fx MVC algorithm shows more robust to impulsive noise and can obtain a low steady-state error.Therefore,the Fx MVC algorithm,which is set to a small step size,is used as a slow filter.To further improve the algorithm's suppression of impulsive noise,the hyperbolic tangent function is used to bound the amplitude of the reference signal and the error signal of Filter1,drawing on the idea of limiting the amplitude.The use of the convex combination method reduces the steady-state error while increasing the speed of convergence.Finally,the MATLAB model is used to generate simulated noise data for simulation,so that the effectiveness of the algorithm is initially verified and the superiority of the algorithm compared with other algorithms is demonstrated.In addition to this,a Fxe LMS algorithm based on the Fx LMS algorithm is also proposed in this paper,which uses a Gaussian error function to vary the cost function.The transformed error signal is kept within a certain range in the weight update,reducing the impact of outliers on the stability of the algorithm.The convergence speed is affected by a certain degree of compression of the error signal.In order to improve the convergence speed,the idea of variable step size is introduced and an improved normalized variable step size is used to improve algorithm to obtain the EVSS-Fxe LMS algorithm.The EVSS-Fxe LMS algorithm shows an improvement in convergence speed and guarantees steady-state error.The effectiveness of the proposed algorithm is verified in simulations with simulated noise and in-vehicle noise during speed bumps as the target noise.Finally,the two improved algorithms are experimentally validated by using hardware-in-the-loop test platform of the active noise control system.The noise reduction effectiveness is tested by employing simulated noise and collected real invehicle noise as the target noise,respectively.The results of the experiment are analyzed,and the conclusion is drawn that the active noise control system equipped with the two algorithms proposed has a significant noise reduction effect on both the simulated noise and the real in-vehicle noise.The analysis for the frequency noise reduction effect of the steady state noise shows that both the C-Fx MVC and EVSS-Fxe LMS algorithms have a noise reduction effect for noise with the frequency range of 80-300 Hz.The average noise reduction achieved by both the C-Fx MVC and EVSS-Fxe LMS algorithms is up to 4 d B for the simulated light impulsive noise.When the target noise is real in-vehicle impulsive noise that generated by crossing a narrow speed bump at35km/h speed,the noise amplitude after noise reduction by the proposed algorithms is significantly attenuated at the speed bump,and the average noise reduction can be also up to 4 d B.