The spatial freedom and the unconstrained work-space of Multi-Rotor Aerial Vehicles (MRAVs) are promising characteristics for many applications that require physical interaction with an environment, such as contact-based inspection, non-destructive testing, and human assistance in remote areas. However, Aerial Physical Interaction (APhI) control remains a major challenge considering the high non-linearity of MRAVs, their inherent instability, and the limited capabilities of the actuators. Classical interaction control methods, which include hybrid position/force control, impedance, and admittance control, are limited by their reactive nature, where the control action is not optimized for the future horizon, and their ad hoc solutions to satisfy the system constraints.
Model Predictive Control (MPC) exploits the a priori knowledge of the system dynamics to predict its behaviour in the future horizon and optimize the control action to achieve the control objective, while the system constraints are included in the optimization problem. In the literature, MPC-based APhI controllers are proposed using a hybrid position/force control approach, without any consideration of impedance control, or cascaded control architectures. In this thesis, the use of Nonlinear Model Predictive Control (NMPC) for APhI control is investigated in three configurations, impedance control, admittance control in a cascaded architecture, and hybrid position/force control.
Three NMPC-based control approaches, exploring the three configurations, are proposed, implemented, analysed, validated with real-time simulations of interaction tasks with different environments.
To join the presentation via Microsoft Teams click here