Active disturbance rejection twice optimal control for time delay system

Abstract

Time delay makes system control be a challenge. Twice optimal algorithm is an effective approach to address time delay. However, it needs model information of a system. In order to reduce its dependence on system model, active disturbance rejection control is adopted. With the help of extended state observer, uncertainties and disturbances are able to be estimated, and those uncertainties and disturbances can be compensated by control signal in real time. By combining active disturbance rejection control and twice optimal control, active disturbance rejection twice optimal control (ADRTOC) is proposed. Comparisons between ADRTOC and PI plus Smith predictor control have been presented. Numerical results confirm that ADRTOC has better performance.