Adaptive Visual Servoing Control of Closed-Architecture Robots With Both Visibility Constraints and Tracking Error Constraints

Abstract

Although there are significant achievements in adaptive visual servoing control of robots with the consideration of visibility constraints or tracking error constraints, most of these approaches are limited to addressing one aspect of the two constraint issues. This article presents a novel adaptive visual servoing control solution specifically designed for robots with closed architecture, effectively managing both visibility constraints and tracking error constraints simultaneously. First, by appropriately combining the image-space position-based constraining function and tracking error-based constraining function, we develop a novel composite barrier function (CBF), enabling the transformation of the visual servoing system with both visibility constraints and tracking error constraints into a functionally equivalent “unconstrained” one. By stabilizing the “unconstrained” system, both visibility constraints and tracking error constraints are handled simultaneously for the first time. Then, we introduce a CBF-based adaptive visual servoing loop controller, designed to dynamically compute the position or velocity commands for the torque control loop. The controller's proven ability to guarantee the convergence of the visual servoing system, without modifying the manufacturer-embedded torque controller, renders it suitable for a wide range of industrial/commercial robots with closed architecture. Experiments are conducted on an industrial manipulator to exhibit the superior capabilities of the method.