PhD project title and outline, including interdisciplinary dimension:
Human-robot motor coordination: robots inferring human intention
The role of robots in our society is ever-increasing. Robots are devised to autonomously perform many tasks currently performed by humans. In designing robots, situations involving human-robot collaboration – very likely to occur for instance on factory floors that cannot yet be completely automated – specific challenges remain. The control of robot movement must be finely coordinated with the movements of the human counterparts. Both the robot and human likely infer intentions from the observed movements of their counterparts. Indeed, humans can infer much about another biological entity from rudimentary features of its movements (called biological motion). For human-robot interactions, additional feedback systems could inform the human about a robot’s upcoming (intended) movements, since these intentions can be accessed. The reverse is not true: a robot does not have access to the movement planning stages of the human, which occur inside the brain. For a robot to infer the intentions of their human counterpart, reliance on principles of biological motion thus seems essential. Indeed, human movements – particularly in constrained tasks – are relatively stereotypical, which may help in predicting upcoming (intended) movements. This project aims to use modern computational intelligence methods (machine learning, deep learning) to predict human (arm) movements based on features extracted from vision (e.g., optical motion tracking, video cameras) with the specific aim to improve performance of a constrained task that requires human-robot motor coordination (i.e., cooperation). Where possible and needed, the algorithms will be constrained or informed by knowledge about the human ability to extract intention from biological motion. This project is interdisciplinary at its core, given the key role of computer science, engineering, motor neuroscience, and psychology. This is strongly reflected in the supervisory team.
Primary supervisor: Professor Seán McLoone (EEECS)
Secondary supervisor: Dr Joost C. Dessing (Psychology)
third supervisor: Dr Vien Ngo (EEECS)
External Partner/Organisation: Irish Manufacturing Research Ltd.