Real-time integration of sensing, planning and control in robotic work-cells

Abstract

This paper addresses real-time planning and control for robot manipulators in a re-configurable work-cell or an unknown workspace. The considered tasks in robotic work-cell are (1) to catch a moving object by using vision-guidance system, and (2) to follow a path on an unknown surface. These two tasks are commonly performed in the applications of material handling and processing. In order to efficiently and successfully accomplish the tasks, novel strategies for integrating real-time planning and control are proposed to improve agility and reconfigurability of robotic systems and to deal with the uncertainties in the environment by means of multi-sensor fusion. In the proposed schemes, a novel calibration-free stereo vision algorithm and a new hybrid position/force control strategy are utilized. In order to achieve intelligent robot manipulation in an unstructured workspace and reconfigurable work-cell, complementary sensory data from different types of sensors, such as force/torque, encoders and visual sensors, are integrated simultaneously during robot constrained motion. © 2003 Elsevier Ltd. All rights reserved.