Visual Servoing Control for Robot Manipulators

Visual servoing has emerged as a promising approach for controlling robotic manipulators by leveraging visual feedback to enhance precision and adaptability. Despite significant progress, several challenges persist, including robustness to image noise, computational efficiency, and the ability to cope with dynamic environments. Visual servoing techniques are especially effective when the camera and the end effector have a limited workspace. This paper explores the use of visual servoing control in robotic systems, in particular, the comparison between imaged-based and pose-based techniques. Further, these techniques are adapted with the use of an adaptive gain in the control law to improve robustness and convergence. Experimental results for all variants are obtained from implementation on a 6DOF robotic manipulator, using an AprilTag fiducial marker.