Shared Control

• Repetitive manual tasks in controlled environment such as manufacturing process have been partially replaced by robot automation. However, most of the tasks performed by human are still challenging to  be completely automated with robots. It is still difficult to use robots in dynamic and unstructured environments. In addition, tasks that are highly dangerous to humans, such as surgery or autonomous driving, cannot be fully automated until the autonomous agents satisfy extremely high reliability and robustness. It is expected that these tasks will take a considerable amount of time to be replaced by robots.

• We are researching the shared autonomy or the shared teleoperation control that can be applied in such environments. The shared controllers could be useful in the situation where the human operator have difficulties to perform a certain task due to the absence of sensory, perceptual, or motory capability from the remote site. The example of tasks can be tactile-based manipulation, polishing of the curved surface, depth perception, etc. The shared teleoperation control can be potentially applicable to the fields such as surgery, disaster/rescue, nuclear power plants, and space exploration.

• We are currently addressing the following issues:

- Design of shared teleoperation control architecture

- Collaborative pose correction in teleoperative precision insertion task

- Learning operator's intention in teleoperative precision insertion task

- Semi-autonomous control protocols and applied machine learning algorithm to improve the safety and efficacy of robotic super-microsurgery

- Semi-autonomous controller to be used during teleoperation with intermittent time delay