Active noise attenuation using adaptive Model Predictive Control

The Model Predictive Control (MPC) approach based on an adaptive linear model is explored for design of Active Noise Control (ANC) of a one-dimensional acoustic duct system. The adaptive linear model is obtained through the Least Mean Square method and then used as the prediction model in the MPC algorithm. One specific MPC algorithm called Dynamic Matrix Control is designed and implemented for the ANC purpose. The numerical simulation showed the efficiency of the proposed method in dealing with possible changing noise environment and system nonlinearity as well, e.g., the time delay due to the acoustic propagation.