Robot Task Space Analyzer System Calibration

Sensors such as cameras and range finders provide flexibility and safety to telerobotic manipulations. However, in order to use their supportive features, sensor-based telerobotic control requires that both the sensors and the robot arm be calibrated so that the manipulator can target and articulate objects in the environment easily and effectively. The scope of this paper is to define the sensor-manipulator calibration issue, outline the system errors, describe the calibration methods, and explain the algorithms necessary to provide compensation. The problem was divided into three steps: acquisition of calibration data of the sensor and the arm, numerical analysis and determination of needed parameters, and validation of obtained results. Since the task is to find parameters that represent the object position accurately, the problem is treated as multidimensional least squares optimization problem with simple bounds. The function to minimize is the distance between the target and its calculated position. Two approaches are applied to represent kinematics parameters: Euler angles and quaternions. Solutions are analyzed to determine the best method for future planned online calibration checks as well as to perform teleoperation tasks on the calibrated system