A Framework for Visual Servoing

摘要

We consider typical manipulation tasks in terms of a service robot framework. Given a task at hand, such as "Pick up the cup from the dinner table", we present a number of different visual systems required to accomplish the task. A standard robot platform with a PUMA560 on the top is used for experimental evaluation. The classical approach-align-grasp idea is used to design a manipulation system. Here, both visual and tactile feedback is used to accomplish the given task. In terms of image processing, we start by a recognition system which provides a 2D estimate of the object position in the image. Thereafter, a 2D tracking system is presented and used to maintain the object in the field of view during an approach stage. For the alignment stage, two systems are available. The first is a model based tracking system that estimates the complete pose/velocity of the object. The second system is based on corner matching and estimates homography between two images. In terms of tactile feedback, we present a grasping system that, at this stage, performs power grasps. The main objective here is to compensate for minor errors in object position/orientation estimate caused by the vision system.