A Comparative Study of Eye-In-Hand Image-Based Visual Servoing: Stereo vs. Mono

Abstract

Visual control of manipulators provides significant advantages when working with targets with unknown positions. Among all visual servoing categories, the stereo visual servoing has shown to be very effective when dealing with unstructured environments. In this paper a comparative study of eye-in-hand image-based visual servoing (IBVS) for two approaches of stereo and mono vision is presented and various cases of control schemes, and tracking and prediction algorithms are studied. In this study the vision systems are considered to be mounted on the end-effector of a 6 degrees of freedom (DOF) manipulator robot. Additionally, a method for position prediction and trajectory estimation of the moving target in order to use in a real-time catching task is proposed and developed using Kalman Filter and Extended Kalman Filter (EKF) as the trajectory estimators. Using the proposed estimation methods, the quality of the visual servoing in a catching procedure using a 6-DOF manipulator robot is compared for mono and stereo visual servoing systems. Finally, by applying the newly introduced acceleration command-based controller (AIBVS) to the visual servoing system, the results for both cases are presented to compare the effects on the quality of servoing tasks. The aforementioned scenarios of visual servoing are simulated and implemented based on a 6-DOF DENSO 6242G robot.