Ability-based assistance in complementary task planning for personalized human-robot collaboration

Understanding how assistance works is crucial for allowing humans and robots to collaborate as a team. This paper proposes ability-based assistance, in which assistive actions provide resources and information or adjust the context, thereby affecting the task’s required ability or recontextualizing the actor’s ability, thus enabling task completion. This concept is implemented in Planning Domain Definition Language (PDDL) and applied in task planning for collaborative assembly. The demonstration using four personas, including disabled and injured ones, shows that the framework is able to integrate assistance into task allocation and generate an optimal plan tailored to each individual’s abilities.