The Research And Design Of Autocar Active Noise Contro System Based On LMS Algorithm

With the vigorous development of automobile industry, car quality gradually attracts attention by consumers, which allows the car’s noise control levels become a major automobile manufacturer’s problem. The car has a complex noise component, which is mainly composed of low frequency noise, and the noise portion sandwiching an intermediate frequency. Passive noise reduction technology is good at high frequency noise, but at the low-frequency sound field has little influence. Active noise reduction technology makes use of sound waves to cancel each other out and interference principle, the secondary sound waves and original noise of equal amplitude, phase inversion, thereby offsetting each other noise.First at all, the papper introduces the background of this topic, active and passive noise control technology. And then analyze their basic principle of acoustics. Make the realization of technology with the guidance of the theory. At the same time, make an analysis of the noise from car’s own structure.Secondly, because of the adaptive filter owning the advantages of strong real-time and single structure, this paper mainly use this method. Adaptive filter is mainly composed of two parts, the transversal filter and adaptive correction module. Transversal filter is depending on real-time error from the input. The adaptive algorithm is based on error minimization principle to adjust their own weight to adapt the changing environment noise. At the same time, do some analysis and introduction four types of the communication system model of active noise control,that is feedforward, feedback type, multi-channel and mixed types.At last, in the experimental stage,the papper demonstrates the influence on the length of the transversal filter. Make an experiment on fixed step length and variable step size which are based on dual channel system model. And find the ideal of the experiment model parameter variable step length is much better than traditional model not only in theory, but also in the actual cases. This simulation may be further contributed to practical application. Finally at real voice signal simulation, the experimental results from two aspects of convergence and processing time are satisfactory.