Planning and visual-servoing for robotic manipulators in ROS

Abstract

This article presents a probabilistic road map (PRM) and visual servo control (visual-servoing) based path planning strategy that allows a Motoman HP20D industrial robot to move from an initial positional to a random final position in the presence of fixed obstacles. The process begins with an application of the PRM algorithm to take the robot from an initial position to a point in space where it has a free line of sight to the target, to then apply visual servoing and end up, finally, at the desired position, where an image captured by a camera located at the robot’s end effector matches a reference image, located on the upper surface of a rectangular prismatic object. Algorithms and experiments were developed in simulation, specifically, the visual servo control that includes the dynamic model of the robot and the image sensor subject to realistic lighting were developed in robot operating system (ROS) environment.