Eye-in-Hand Visual Servoing Control of Robot Manipulators Based on an Input Mapping Method

Abstract

In image-based visual servoing (IBVS), parametric uncertainties tend to cause the model inaccuracy and limit the control performance. Considering these uncertainties can be embodied by the output-input data from the visual servoing system, this brief proposes an eye-in-hand visual servoing control (VSC) scheme based on the input mapping method, which directly utilizes the past output-input data to enhance the original feedback control law rather than identifying the model. The system with the input mapping method is proven to not only maintain the stability of the original VSC but also accelerate the convergent rate. The results of the experiments on a manipulator with an eye-in-hand camera demonstrate the superiority of our proposed method.