Visual servoing 2D based on global description: Validation on Koala mobile robot

Abstract

This paper presents the new technique of global visual features based on random process. Using 2D visual servoing algorithm applied on the robotic platform equipped with a wireless camera, the robotic system must always find a desired position stored and referenced by an object and defined in the navigation environment. To operate the assembly formed by a robot + camera, processing unit (computer) which supports the acquisition of the visual information from the imaging device (camera).The processing unit used its data processing result and calculates the control law and sends operation commands to the robot. This technique of global visual servoing based on a stochastic visual information extraction is used to control a mobile robot and it is simulated in the Matlab tool. In fact, this method ensures a reduction of computation time compared to the ancient one.