End-effector control of robots using just remote camera views is difficult due to lack of perceived correspondence between the joysticks and the end-effector coordinate frame. This paper reports the positive effects of Augmented Reality visual cues on operator performance during end-effector controlled teleoperation using only camera views. Our solution is to overlay a color-coded coordinate system on the end-effector of the robot using AR techniques. This mapped and color-coded coordinate system is then directly mapped to similarly color-coded joysticks used for control of both position and orientation. The AR view along with mapped markings on the joystick give the user a clear notion of the effect of their joystick movements on the endeffector of the robot. All camera views display this registered dynamic overlay information on-demand. A preliminary test using fifteen subjects comparing control of performance with and without the coordinate mapping was performed by using a simple insertion task. Preliminary results indicate a significant reduction in distance, reversal errors and mental workload.

Two inverse kinematics algorithms were implemented in a tele-operated robot system and evaluated with a user performance study. The kinematics algorithms were designed such that the point of resolution (POR) of the robot arm's wrist and the end-effector was controlled by joysticks, one each for rotation and translation. Operator performance was evaluated with "peg-in-the-hole" type tasks using both the wrist and end-effector POR modes. Wrist kinematics resulted in faster performance times, however, with longer average distances traveled while the opposite effect was observed with end-effector kinematics. Reversal errors were present equally in both modes, while the end-effector mode showed higher 1-axis use of the joysticks. Implications for remote robotic operation design and kinematics are discussed.

Two inverse kinematics algorithms were implemented in a tele-operated robot system and evaluated with a user performance study. The kinematics algorithms were designed such that the point of resolution (POR) of the robot arm's wrist and the end-effector was controlled by joysticks, one each for rotation and translation. Operator performance was evaluated with "peg-in-the-hole" type tasks using both the wrist and end-effector POR modes. Wrist kinematics resulted in faster performance times, however, with longer average distances traveled while the opposite effect was observed with end-effector kinematics. Reversal errors were present equally in both modes, while the end-effector mode showed higher 1-axis use of the joysticks. Implications for remote robotic operation design and kinematics are discussed.