Paper accepted at Robotics: Science and Systems (RSS 2021)
Congratulations to Chris Mower, Joao Moura and Sethu Vijayakumar
- Chris Mower, Joao Moura and Sethu Vijayakumar, Skill-based Shared Control, Proc. Robotics: Science and Systems (R:SS 2021), [Held Virtually] (2021).
Performing motion patterns - referred to as skills - (e.g. waves, circles) during teleoperation is an integral part of many industrial processes such as concrete spraying, welding, plastering, and cleaning. Maintaining these whilst simultaneously avoiding obstacles and traversing complex terrain requires expert operators. We propose a skill-based shared control framework for incorporating the notion of skill-assistance by estimating the operators intention, and optimizing a predicted trajectory within a computationally feasible MPC loop. Our shared control method leverages a novel parameteization for state and control trajectories based on a geometrical primitive known as Clothoids. We perform several experiments on a hardware mock-up, validating the effectiveness of our method to recognize a switch in the intended skill, improve quality of output motion, and even avoid dynamically changing obstacles whilst respecting the features of the skill.