Assistive obstacle-free path planning in the image space for teleoperation tasks using monocular camera

摘要

To assist operators in precise teleoperation tasks within complex environments, assistance autonomous modules are required. A semi-autonomous path planning module that provides a reliable direction for the operators, is of paramount of importance in this field. This paper proposes a novel assisted method using a fixed camera for guiding the operator to control the manipulator remotely. The method fuses image-based path planning and collision avoidance techniques to achieve an obstacle-free guidance path in the image space. Moreover, this method enables the visualization and projection of the 3D generated path and orientation profile in the 2D image space. It gives the teleoperator a powerful depth cue, as well as the correct pose of the end-effector, to increase the system performance and reduce the operator task error. The experimental results verify the visual framework's capabilities to assist the operator in a teleoperation task.