- Develop a safe adaptive, cognitive deliberation system capable of detecting unexpected situations and provide contingency plans to handle them
- Develop formal tools that ensure correct execution of behaviors and contingency plans instantiated by the cognitive deliberation system both at design time and run time
- Integrate the tools developed in the project in an open-source software toolchain for behavior developers
Results of CONVINCE will be validated on three different real-world use cases, each of which presents technical difficulties and utilizes robotic systems of increasing complexity, in different application domains.
Robot Vacuum Cleaner
The robot's capabilities are limited, but it operates in diverse environments whose specific traits are impossible to foresee completely at the design stage. End-users require vacuum cleaners that improve with time, learn the problematic corners, areas, and objects of their home, to avoid getting stuck.
Two manipulators undertake an assembly task. Parts are available in the environment in no prescribed order and should be assembled based on their shape. Occlusions or difficult lighting conditions may impair the visual system. Some parts may be missing or occluded or restrained by other parts.
Robotic Museum Guide
A humanoid robot is utilized to guide visitors inside a museum, describing the artworks. Part of the computation is performed on-board and part off-board. The robot behavior should change depending on the available functionalities, the status of the environment or the behavior of humans.