Active Noise Control System Design And Implementation On DSP

With the development of the world industrial, various types of noise has been generated,such that noise pollution has become one of the four major pollution. Noise causes more harmto the human body, the noise reduction has then become more important and urgent.Noise control is divided into active noise control (ANC) method and passive noise controlmethod. This thesis studies active noise control algorithms, and the basic principle is to ofsetnoise with a signal which has the same amplitude and opposite phase as the noise.In this thesis, the least mean square based adaptive algorithm is derived and analyzedin terms of performance and convergence conditions. An ANC system is designed, whichis based on the Filtered-X Least Mean Square (FX-LMS). The whole system of the noisecontrol is simulated with practical frequency-domain characteristics in the sub-channel. Thecanceling efect of diferent kinds of noise is analysed, and the influence of the step is studied.Meanwhile, an adaptive step algorithm is proposed to solve the problem of how to choose thestep size.A hardware platform is built for a single-channel noise control system. The core chipused in this system is digital signal processor (DSP) chip which is configurated with the chipclock generator, multi-channel bufer serial port and built-in integrated I2C bus. In addition,it is also configurated with the audio chip to drive the speaker and microphone. According tothe algorithm, the development environment and hardware platform, the process of the wholesystem is designed in software form.Through a large number of simulations and the actual channel characteristics, the de-signed feedforward ANC system can efectively filter the low frequency noise and performswell. Meanwhile, the system is simple and easy to operate.