Adaptive Active Disturbance Rejection Control and Its Simulation Based on BP Neural Network

Because the fixed parameters of the extended state observer(ESO) reduce the estimation precision of "total disturbance" and control effect for the systems which the parameters of the controlled objects change largely and fast or there being serious and uncertain outside disturbance,an adaptive active disturbance rejection controller(ADRC) based on BP neural network was proposed.The significance of introducing adaptive ESO as well as the structure of ESO was analyzed,then the adaptive ESO which parameters are adjusted online by means of BP neural network was applied to ADRC.Simulations show that the improved ADRC has higher estimation precision,smaller range of controlling quantity,high robustness and anti-interference performance compared with conventional ADRC.