A New Adaptive Feedback Neural Controller For Broadband Active Noise Control Systems

This paper develops a new adaptive feedback neural method for broadband active noise control systems. This controller consists of three layers, in which weights of the output layer update for weights of the hidden layer. Hence, the proposed controller only uses an adaptive filter at the output layer with a fast-learning algorithm. The proposed algorithm and computational complexity are analyzed in this paper. Numerical simulations are implemented with the zero-mean white noise signal as a primary noise source. The efficiency of the proposed controller is verified by results comparison with the conventional controller.