Robot manipulator utilizing image based visual servoing

Most of the robots in the industry perform tasks by using a pre-programmed sequence of events. The Pegasus Arm, which forms part of the Computer Integrated Manufacturing (CIM) system, was no exemption. With the old set-up, only a single position was allowed and locators had to be employed to secure the pellet in its right place. Hence, visual servoing is needed. Visual servoing involves the concept of a closed loop position control, wherein feedback can be received from external disturbances such as an abrupt change in the position and orientation of the object as the vision and motion are simultaneously run. There were two types of visual servoing these are position based and image based. Position based uses Cartesian coordinates and seeks for the image error function. The group used one camera that is mounted in the end-effector of Pegasus robot to perform image extraction by designing a filter that will isolate the feature of the blue pellet from its surroundings. Robot controller circuit had been made in order to manipulate the six motors. Robot Toolbox in Mathlab 6.5 was used in control and simulation of the robot arm. Thus, Uncalibrated Visual Servoing algorithm was done in the said environment. With the implementation of image based visual servoing, a more humanlike robot manipulation was achieved, wherein the gripper could position itself in any pellet orientation within the pick-up station even with the absence of locators and under normal room lighting. Also, the robot arm could follow moving object as long as the movement of the pellet is in the camera\u27s range of sight