Towards human capability estimation to enhance human-robot team performance

Abstract

Skilled labor shortage is a prominent challenge in the world of work. Meanwhile, age-related disabilities or injury lead to at least temporary performance limitations, which make people unfit to work. Consequently, even less workers are available. By employing human-robot teams, the performance of these people may be restored. This requires a good artificial understanding of the human’s capabilities, as generic robot behavior is not feasible with the highly individualized manifestations of disability. We present an approach that allows the robot to autonomously assess human capabilities based on standards from occupational medicine. The method does not only indicate the presence/absence of capabilities, but gives them a discrete rating. This allows the robot to better define its own behavior as a mixture of supportive actions based on gaps in the detailed capabilities.