Research On Multi-channel Active Control Of Road Noise Inside Vehicles Based On Intelligent Algorithm

With the improvement of automobile quality,the noise inside the vehicle has been at the heart of evaluating the comfort of the automobile.At present,through in-depth analysis and research on vehicle vibration and noise,engine and transmission system noise has been effectively controlled.The proportion of noise generated by the interaction between tires and the road surface(hereinafter referred to as "road noise")in the vehicle is The increasing influence,reducing road noise in cars,and improving sound quality in cars are important trends in the development of the automotive industry.Road noise is low-frequency wide-band random noise.Traditional passive noise reduction technologies,such as sound absorption,sound insulation,and sealing,are difficult to effectively suppress.Therefore,active noise control technology has aroused the attention of automobile science and technology workers.At the same time,in order to meet the noise reduction needs of multiple locations in the car,multi-channel adaptive active noise reduction technology needs to be studied.This paper studies the basic principle,system structure and algorithm implementation of noise multi-channel active control technology,analyzes the noise characteristics of road noise in the car,and proposes a multi-channel active control strategy of road noise based on neural network technology.The vibration signal at the connection point is extremely correlated with the road noise signal in the car.Therefore,the multi-channel active road noise control strategy in the car can be proposed.A neural network method with good identification and prediction capabilities for time series signals is used to pass the suspension.The vibration acceleration signal is used to identify the road noise signal in the car,and then the multi-channel noise active control algorithm can reduce the road noise.On the basis of the proposed control strategy,the following research work has been completed:First,the control scheme of the multi-channel noise active control system is determined by the way of comparing many algorithms.The secondary channel identification and active noise reduction of the multi-channel noise control system are respectively completed using the LMS algorithm based on the stochastic gradient and the multi-channel FxLMS algorithm.Secondly,based on the vibration and acceleration signals collected at the junction of the vehicle and the vehicle body at the constant speed of the test and the noise signals near the driver's and rear passenger's ears,the low frequency characteristics and multiple correlations were analyzed to establish the interior road.Elman neural network model for noise identification.Then,the multi-channel noise active control system model was established,and it was integrated with the multi-reference LMS algorithm synthesis in-vehicle road noise model and the Elman neural network in-vehicle road noise identification model,respectively,and the existing in-vehicle road noise active models were built respectively.The control technology model and the neural network-based multichannel active control model for in-vehicle road noise are compared to verify the effectiveness and feasibility of the proposed multi-channel active control strategy for in-vehicle road noise.Based on the vibration and noise data collected by the experiment,the two sets of models were simulated and analyzed respectively.The simulation results show that both sets of models have good noise reduction effect in the range of road noise frequency,and can achieve simultaneous noise reduction at multiple points.The noise reduction of existing road noise control technology is mainly distributed within 10 dB,and the peak can reach around 20 dB.The noise reduction effect of the in-vehicle road noise active control system based on neural network identification is not inferior to the existing in-vehicle road noise multi-channel active control strategy,even the peak noise in the 0-50 Hz range is better,and can reach about 25 dB.Finally,based on the proposed in-vehicle road noise multi-channel active control strategy and the in-vehicle road noise multi-channel active control Simulink model,a hardware-in-the-loop simulation platform for the in-vehicle road noise multi-channel active control system was built,Based on the data,a hardware-in-the-loop simulation test was carried out on the built system.The results show that the neural network-based multi-channel active control system for road noise has a certain noise reduction effect in the frequency range of 20-100 Hz,and can achieve multiple points simultaneously.Noise reduction,the amount of noise reduction is mainly distributed between 2dB-8dB,and the noise reduction effect at the peak of the noise is better.