Robust control strategy for multirotors independent of varying equipment configuration

Abstract

Modifications of multirotor UAVs (unmanned aerial vehicles) are increasingly popular by integrating tools, sensors, or robotic arms depending on types of applications. In these circumstances, precise motion control against model uncertainty and disturbance is essen- tial. This dissertation presents a robust motion controller for a multirotor combined with additional objects. By conducting thorough analysis on the dynamics of the objects-added multirotor, the robust motion controller which consists of two loops is proposed. The inner- loop controller is designed based on disturbance observer (DOB) to recover the dynamics of the modified multirotor similar to the bare multirotor. Such recovery allows to design the outer-loop controller based on a typical multirotor dynamic equation. In addition, in the process of the inner-loop controller design and the associated stability proof, the safe operation bound of the modified multirotor is revealed in accordance with the amount of inertia added to the multirotor. To effectively address the performance of the robust controller, cooperative aerial ma- nipulation is performed with two aerial manipulators which are multirotors integrated with multi-DOF robotic arms. When two aerial manipulators are physically connected through an object, each multirotor would experience disturbances generated by the robotic arm, the object, and the peer aerial manipulator. In this situation, by using the proposed robust controller, it is possible to control the multirotors precisely regardless of those undesirable effects. Hence, each aerial manipulator can be handled in a decentralized manner rather than incorporating the complicated entire dynamics including multiple aerial manipulators and the object. Furthermore, a motion planner that generates the reference velocity of the aerial manipulators is proposed for manipulating an object cooperatively. To assure safety during the cooperative manipulation, the internal force between the aerial manipulators is estimated and controlled. Also, unilateral constraints including the collision avoidance between the object and aerial manipulators are considered. All the designed features are prioritized and transformed into the reference velocity of the aerial manipulators. In this dissertation, all the proposed control and guidance laws are validated with suc- cessful experiments conducted with custom-built multirotors combined with multi-DOF robotic arms. First, the motion control performance is evaluated with a trajectory tracking test conducted with the multirotor and the four-DOF robotic arm in motion. Moreover, an autonomous cooperative manipulation is performed with two multirotors with three-DOF robotic arms. The experimental setup and results are presented with the detailed hardware description. The proposed double-loop controller with verifiable robustness and systematic consid- eration of various factors for safe cooperative manipulation can contribute to the practical application of multirotor UAVs.