Simulation and experiments for a vision-based control of a 6R robot

In this paper, we present both the simulation and experimental results for position-based visual servoing control of a 6R robot using two stationary cameras . This method can deal with real-time changes in the relative position of the target with respect to the robot, while providing greater accuracy. In addition, the servo control structure is independent of the target pose coordinates. Direct and inverse kinematics equations, in addition to dynamic equations of the 6R robot, have been derived and simulated. Then we prescribe simulation of image processing, object recognition and pose estimation for the end effector and target-object in 3D Cartesian space and visual control of the robot. Afterwards, performance tests of 6R robot using two cameras have been implemented. Finally, experimental results obtained from actual implementation of visual control and tests of 6R robot in lab are presented. Analysis of error and test data has been carried out according to ISO9283, ANSI-RIA R15.05–2 standards and the MATLAB statistical toolbox. An efficient algorithm is designed for collision detection and contact determination between the 6R robot links and its environment during their rotation in controlling and performance tests. In this technique, bounding spheres on different parts of 6R robot are used to detect collisions among them.