Toward practical semi-autonomous teleoperation: Do what i intend, not what i do

Abstract

The discussed research trajectory could begin improving the capabilities and robustness of teleoperation immediately. Initial work related to the RRM project described in Section I-A is beginning to incorporate explicit task models that can be used to bootstrap more sophisticated approaches like those used in the Language of Surgery project. The inclusion of task and skill modeling into robotic automation, and in general the movement to more seamless human-robotic task collaboration, will have long term effects in both the technical aspects of robotics, as well as industrial and societal acceptance of robotics. One of the significant limitations to the continuing adoption of robotics has been a lack of trust. This lack of trust often manifests itself in a technical form; for instance, engineers in industry will often double check by hand the calculations or task execution plans generated by an autonomous system before allowing an action to be performed. This gives us greater confidence that the robot is performing as we would, given the situation.