Laser assisted telerobotic control for remote manipulation activities

The effort in this work has been to innovatively use range information from a laser sensor mounted on the end effector of a remote robotic arm in a telerobotic system to assist the user in carrying out remote tasks in unstructured environments. Assistance is provided in the form of Traded Supervisory Control where the human is involved in high level activities such as decision making and the machine generates task plans and executes tasks autonomously using laser data and machine intelligence. In this way human planning and high level decision making capabilities are combined with machine computational and precision task execution capabilities in an optimal way. Laser range data has been used in a novel way to generate trajectories and virtual constraints that assist the user either by executing trajectories autonomously or by guiding the user in teleoperation along specific virtual constraints. The ability of the laser to generate path plans and execute them autonomously, and generate 3D geometry information is another novel feature of the project. This has been achieved without using sophisticated sensory suite and extensive computer processing. The user simply points to certain locations in the unstructured environment by teleoperating the remote arm using the master arm and presses certain keys on the keyboard. The machine using laser data and its intelligence generates the appropriate trajectories, virtual geometric surfaces and path plans which assist the user in executing the task. Time and accuracy results in executing a remote manipulation tasks on a real-time telerobotic system with master and slave arms, with and without laser based assistance have been generated and compared to validate the hypothesis that laser based assistance improves task performance and reduces the cognitive load on the user. To improve dexterity of the arm and to enable smooth and stable control of the arm, singularity avoidance techniques have been implemented and results in simulation have been presented. Accuracy results to validate the motion control algorithms of the robot by comparing trajectories in simulation and on the robot have been generated