作者: Judith Müller , Udo Frese , Thomas Röfer , Rodolphe Gelin , Alexandre Mazel
DOI: 10.1007/978-3-319-02934-4_9
关键词: Artificial intelligence 、 Motion (physics) 、 Robot 、 Stereopsis 、 Computer vision 、 Whole systems 、 Object (computer science) 、 Detector 、 Nao robot 、 GRASP 、 Computer science
摘要: In this paper we introduce an online object manipulation system for the NAO robot that is able to detect and grasp out of a human hand then give it back in real-time. Known objects are rendered from 3D models detected stereo contour-based by using new vision head NAO. order objects, motion trajectories generated A* planner while avoiding obstacles. safely release into hand, combination tactile force sensors carrying arm used whether someone touched grasped object. We performed quantitative experiments evaluate quality detector, time object, as well number successful grasps. demonstrated whole on real robot.