A semi-autonomous vision-based navigation system for a mobile robotic vehicle

Abstract

Direct teleoperation of a mobile vehicle over a communications channel with delay on the order of one second is problematic for precise maneuvering and obstacle avoidance. To curb both operator vigilance and maneuvering mishaps during vehicle teleoperation, this thesis aims to develop a semi-autonomous vision-based navigation system for a small mobile robotic vehicle designed by Draper Laboratory. The system relies on monocular vision processing to employ an optical flow balance strategy for real-time obstacle avoidance. The system utilizes a multimedia processor for video processing and is implemented within a payload module built for a visual servoing task. The system design relies on a flexible communication protocol to integrate with the robot and operator control unit. Overall, the system is a useful augmentation for a flexible robotic platform for enabling better reconnaissance.