A method of active noise control using adaptive neural networks with improved robustness

The time-variance of plant parameters, which is not considered usually by traditional control methods, is one of the main problems in the application of active noise control. This paper firstly introduces an extended plant adaptive neural network model which can be on-line adapted conveniently, then presents a method of active noise control using adaptive neural networks. With separately adapting of the control network and the model network in control, the problem of time-variance of plant parameters is solved, and the robustness of whole system is improved obviously. It is showed by experimental results that this method has a good robust stability for abrupt disturbances of the plant.