The Research And Design Of Automotive Active Noise Control System Based On Intelligent Algorithm

Active noise control system mainly uses an artificially generated noise offset to stackwith the original noise so that the original noise will be reduced. With the development ofautomobile industry, the attention for the ride comfort has made automobile noise control becomean important research direction. Automobile noise control aims to reduce the automobilenoise which will cause damage to passengers and instruments in the car.A car is a complex noise environment whose noise is mainly composed of lowfrequency noise, and mixed with some middle frequency noise. Traditional passive noisecontrol methods, for example, muffler materials, have good soundproof effects to high frequencynoise, but for low frequency noise which is a main constituent of automobile noise, these methodsare not very efficient. Active noise control is based on the interference cancellation principle ofsound waves, by using artificially generated secondary noise to reduce the original noiseand achieve the purpose of noise reduction.The current active noise control systems mainly adopt adaptive filter method. They can bedivided into two parts according to the function of the structure, which are the adaptive algorithmand the transversal filter. The adaptive algorithm is in charge of weights adjustment ofthe transversal filter in accordance with the real time error inputs so that the transversal filter canbe adaptive to the change of the environment.This article mainly studies and designs the automobile active noise control system andits fault diagnosis system. The main work of this paper is as follows:(1) Design of the automobile active noise control system. This system adopts the mechanismof online identification of LMS filter and introduces variable step sizes to eliminate the vehiclenoise in real time. Here, using simulation to verify the method. With10order,50order and100order FIR filter respectively, and a0.0001step size. Simulation results show that, with themechanism of on-line identification and variable step sizes, the LMS filter can be a good way to eliminate noise interference and extract the useful signals. And the filter effect is better with theorder increasing.(2) Design of sensor fault diagnosis system of the above active noise control system. Thediagnosis system mainly uses a two-layer system model which consists of a SVM model andseveral RBF neural network models. On the basis of the prediction function of those models, thefunctions of real-time sensor fault diagnosis, faulty sensor localization and signal reconstructionof the faulty sensor in the active noise control system are realized. Here, using simulation toverify the method. Simulation results show that the system could check, locating and reconstructhe signal in about3seconds. Simulation results also show that the fault diagnosis system of thistwo-layer model is efficient in fault diagnosis, localization as well as signal reconstruction.This study takes the automobile cab noise as the control target, and designs the active noise control system based on adaptive filter in order to realize noise control. Simulation experiments results indicate that this the active control system has a good noise reduction effect. On the basis ofthe above work, we also design a fault diagnosis system based on double layer intelligent algorithm model for this adaptive active noise control system. Simulation results also show that the faultdiagnosis system can realize fault diagnosis, fault localization as well as signal reconstruction in the case of single sensor fault.