Object manipulation with aerial robots is a highly researched topic with a great potential impact. Aerial robotic manipulators can operate in a large workspace without being hindered by uneven terrains. However, aerial robots suffer from payload and flight endurance limitations. A multi-robot approach help solve these issues but poses interesting research challenges on how to achieve the robots’ cooperation. As manipulation tools, cables have often been chosen in this context as they are lightweight and economically advantageous.
In [1], the authors study the manipulation of a cable-suspended object using two aerial robots; they show that there is a free parameter, namely the internal force induced by the robots on the manipulated object, which influences, besides the robots’ control effort, the sensitivity of the system to parameter uncertainties. This is especially relevant as parameter uncertainties are never completely avoided in practice.
This assignment, supervised in collaboration with Dr. Tognon from Inria Center of University of Rennes (France), aims to answer the following research question:
How to exploit the aforementioned sensitivity information to let the system move in a known cluttered environment in the presence of bounded uncertainties on the system’s parameters?
Main steps of the assignment:
- Study of the provided literature by the supervisors
- Planning of the system’s references to let it perform an obstacle-free trajectory in a cluttered environment:
- In the first phase, the trajectory of the manipulated object can be assumed known and the planner chooses the internal force to optimize cost functions such as energy consumption with the worst-case obstacle avoidance constraint.
- In the later phase, the planner also optimized the object’s trajectory and not only the internal force.
- Experimental validation of the results on real robots is a plus and will be considered if the time allows.
Working simulators of the studied system are available, both in Matlab-Simulink and in Gazebo interfaced with Matlab-Simulink.
Candidate’s profile: The ideal candidate is motivated and independent, and has a research-oriented mindset. They have knowledge in at least some topics among Automatic Control, System Modeling, Optimization, and willingness to learn.
Supervisors: Dr. Chiara Gabellieri, Prof. Antonio Franchi, External Supervisor: Dr. Marco Tognon
How to apply: Send an email to Dr. Chiara Gabellieri <c.gabellieri@utwente.nl> with Prof. Franchi, <a.franchi@utwente.nl> in cc. Please include in the email a brief motivation letter, your CV, and exam records.
References:
[1] Gabellieri, C., Tognon, M., Sanalitro, D., & Franchi, A. (2023). Equilibria, Stability, and Sensitivity for the Aerial Suspended Beam Robotic System Subject to Parameter Uncertainty. IEEE Transactions on Robotics.