Robotic Goal-Based Semi-Autonomous Algorithms Improve Remote Operator Performance

摘要

The focus of this research was to determine if reliable goal-based semi-autonomous algorithms are able to improve remote operator performance or not. Two semi-autonomous algorithms were examined: visual servoing and visual dead reckoning. Visual servoing uses computer vision techniques to generate movement commands while using internal properties of the camera combined with sensor data that tell the robot its current position based on its previous position. This research shows that the semi-autonomous algorithms developed increased performance in a measurable way. An analysis of tracking algorithms for visual servoing was conducted and tracking algorithms were enhanced to make them as robust as possible. The developed algorithms were implemented on a currently fielded military robot and a human-in-the-loop experiment was conducted to measure performance.