An active disturbance rejection approach based on signal processing

A digital signal processing approach to estimate and reject disturbances under an active disturbance rejection control (ADRC) methodology is presented. In this paper a technique to assess the current values of the disturbance based on the previous input and output signals is developed. The on-line disturbance estimation is made through the use of a polynomial least squares regression. An excess in the number of data over the order of the polynomial regression is proposed in order to obtain a low pass filter. This estimation allows to simplify the control design in to a linear one. Corresponding performance, stability and robustness are analyzed. The propose approach is tested via numerical simulation in a trajectory tracking task under perturbed linear and nonlinear differentially flat systems. The simulations showed that the technique presented to estimate disturbances is successful in its application to active disturbance rejection control on linear and nonlinear systems.