| KEYWORD |
Data-Driven Control of Soft Robots
Thesis in external company
Reference persons FABRIZIO DABBENE
Description Traditionally, the term robots has referred to devices comprised of rigid links actuated through motorized joints, capable of executing predefined task-specific operations. This conventional definition inherently constrained the application of robots to contexts wherein their inflexible structure did not constitute a significant obstacle. However, the escalating demand for devices suited to tasks incompatible with the conventional rigid robot structure has prompted a reevaluation and redefinition of the original concept of a robot.
Recent years have witnessed a remarkable shift in research focus towards the development of biologically-inspired robotic system. These systems, commonly known as “soft robots”, integrate elastic elements into conventional rigid-body structures, striving to achieve performance levels akin to those observed in the natural world.
Unlike other application domains, in soft robotics, the early endeavors of controller design have predominantly relied on machine-learning strategies rather than model- based techniques. Nevertheless, providing analytical guarantees of perfor mance with these approaches remains a well-known open challenge. Therefore, recently, it has been observed a prolific interest in inverting this trend, with an increased focus on designing model-based control strategies suited to effectively deal with soft robots.
This series of theses is aimed at analyzing the improvements achievable by the synergic combination of model- based and model-free data-driven control approaches. The thesis will be carried on in the context of a MUR funded project on assistive and rehabilitative robotics.
See also cnr - thesis proposal on data-driven control on soft robotics.pdf
Deadline 17/11/2024
PROPONI LA TUA CANDIDATURA