Active disturbance rejection by state feedback: Experimental validation in a 3-DOF quadrotor platform

The central aim of ADRC is to actively reject the disturbance in the absence of a detailed mathematical model of the plant or the disturbance itself. For that purpose, the controller is designed based on the ideal, disturbance free, process where any discrepancies between the actual process and the ideal one are considered as disturbances, being estimated and, afterwards, canceled. The main contribution of this paper is twofold. First, all the process information is used as an input to the enlarged state observer, to get a better estimation of the disturbances. Then, the approach is applied to control pitch and yaw variables in a lab quadrotor, with some experimental results showing the improvements with respect to previous controllers.