Active Noise Control Based On Related Signal Separation Methods

With the development of modern industry, the noise pollution is much more serious,Traditional noise control methods are better used in the high-frequency noise control, butthe effect of low-frequency control is not obvious. Active noise control methods areproposed to solve this problem. Because of its small size, light weight, easy to control,many experts and scholars put a good deal of effort on the study of active noise control.Active noise control technology is basically based on the linear adaptive filter theory,that is, by using different adaptive control algorithm to reduce noise. However, suchalgorithms are contradictory on the convergence rate and steady-state offset. Increasing thestep size can accelerate the convergence speed, but it also can increase the steady-stateoffset. Usually such adaptive algorithm can achieve better noise reduction effect for singlechannel, but difficult to achieve the same effect on multi-channel noise reduction. Becauseof the huge computation, slow convergence, and difficult to provide the number andposition of the secondary sound source, make it difficult to achieve remarkable resultswhen using the methods on multi-channel noise reduction.In order to improve the multi-channel noise reduction control methods, we propose touse the blind signal separation method to control the active noises. We use two noisesource signal model for example, and we focuses on the case of the two noise sourcesignal parsing to indicate the blind signal separation method. Simulation results prove thefeasibility and accuracy of the signal separation in the noise reduction model. Two of thesignals that obtained by the completion of the blind separation are two noise signals.Fourier transform was carried out by using the two signals. Then we use these majorfrequencies component information what the noise signal spectrum was shown to completethe secondary sound source signals structure.The difference between the two methods, the based on the blind signal separation ofactive noise control method and the traditional noise reduction methods is that the formalmethods abandoned the traditional adaptive control methods, but to choose the blind signalseparation technology to achieve the mixed sound source signals separation. Because theblind separation technology used for multi-channel noise reduction, you can get the exactinformation of each noise signal.So that each channel of the multi-channel signal can be seen as a single channel to control. We can know exactly the number and location of thearrangement of secondary sound sources. With the traditional noise reduction methods, thisis one of the innovations of this article. Finally, the experiment proved that the methoddiscussed in this paper can be used in a more wide area (L:287.5cm, W:175cm) to achieveeffective noise reduction (average3.5dB), in the case of two-channel noise reduction,comparing with the traditional active noise reduction method.