• mpc
• physical
• interaction
• control
• path
• following
• flexible
• environment
• door
• quadcopter
• drone
• dynamics
• optimization
• problem

In the field of aerial robots, physical interaction with the environment is increasingly investigated, spanning a variety of applications, including inspection, monitoring, and maintenance of infrastructure. Nevertheless, such approaches are restricted to human-made environments, which are characterized by rigid and almost ideal structures, such as walls, ceilings, or pipes. In this project, we want to develop a solution for enhancing aerial physical interaction with flexible structures, which can enable a wider set of applications, e.g., maintenance of power lines, data collection within tree canopies. First, we read some literature and got familiar with state-of-the-art solutions for aerial physical interaction, understanding the current limitations and the requirements necessary for interacting with compliant environments. Therefore we formalized and implemented an optimal control strategy through MPC for safe and robust physical interaction, in order to perform pushing and sliding tasks on flexible structures. We developed a simulated environment (made by a door with a spring hinge) for debugging and validating our approach. Finally, we tested and validated the entire framework on a flying platform (a quadcopter with an arch shell for the interaction) through several experiments.