Model based task programming for secure multimodal human-robot interaction

Abstract

This note presents a preliminary study on a high-level model based task programming framework for secure multimodal human-robot interaction in industrial context. The case study of robotic polishing tasks, in particular the task of polishing the whole surfaces of objects of interest, is considered. Our main focus is on system modeling and optimal synthesis of surface coverage trajectories in robotic polishing tasks. Two heuristics are proposed to solve the optimal synthesis problem approximately, based on different considerations and assumptions. We then briefly discuss about polishing tasks other than coverage problem that are specifiable in the language of linear temporal logic and state space reduction heuristics based on learning from expert operators. Finally, we introduce a design of the specification input device for the task specification module.