Thesis Project Form

When two or more than two people participate in a physically interactive task, they are mutually aware of who is participating in the collaborative activity and can seamlessly adapt to the state of the other, for instance slowing down when the partner is tired or increasing the effort if the partner is weaker. The same capability of understanding the other is used in competitive tasks, to ensure the selection of the best strategy to win the competition. Being aware of the partner’s state would be crucial also for robots aiming to support humans in a physical collaboration, or to appropriately challenge them in a training task.

This thesis work aims to integrate existing frameworks for the detection of human inner state and behaviour with the use of sensorised objects and proactive haptic devices, possibly improving their performance by adding additional features (e.g., style of the action [1], emotional state, and level of engagement) and using this information in adaptation to the collaborative activities. Experiments with recruited participants can use existing and available sensors and devices (Wristbot [2], iCube [3]) and will assess the framework's capacity and its impact on the quality of interaction in collaborative scenarios.

Proposed work plan and expected results
1. Understand, run and evaluate collaborative activities with the support of haptic device such as the Wristbot.
2. Understand, run and evaluate collaborative activities with the support of sensorised objects such as the iCube.
3. Run an experiment to assess people's behaviour in collaborative tasks.
4. Validate if the intervention of the artificial agent in support to the collaborative activity optimally adapts to the group of individuals in order to achieve shared goals.
The thesis may also produce a scientific article to be submitted to an international conference or journal (the students may contribute to the writing of the article, if they want, but this is not mandatory).