Grasp Learning by Active Experimentation Using Continuous B-Spline Model
In this chapter we present a self-valuing learning system based on continuous B-spline model which is capable of learning how to grasp unfamiliar objects and generalize the learned abilities. The learning system consists of two learners which distinguish between local and global grasping criteria. The local criteria are not object specific while the global criteria cover physical properties of each object. The system is self-valuing, i.e. rates its actions by evaluating sensory information and the use of image processing techniques. An experimental setup consisting of a PUMA-260 manipulator, equipped with hand-camera and force/torque sensor, was used to test this scheme. The system has shown the ability to grasp a wide range of objects and to apply previously learned knowledge to new objects.
KeywordsLearning System Local Criterion Action Execution Orientation Learner Unfamiliar Object
Unable to display preview. Download preview PDF.
- 1.A. Bicchi and V. Kumar. Robotic grasping and contact: A review. In Proceedings of the IEEE International Conference on Robotics and Automation, 2000.Google Scholar
- 2.G. Smith, E. Lee, K. Goldberg, K. Boehringer, and J. Craig. Computing parallel-jaw grips. In Proceedings of the IEEE International Conference on Robotics and Automation, pages 1897–1903, 1999.Google Scholar
- 3.A. Morales, P. J. Sanz G. Recatalâ, and Angel Pdel Pobil. Heuristic visionbased computation of planar antipodal grasps on unknown objects. In Proceedings of the IEEE International Conference on Robotics and Automation, 2001.Google Scholar
- 5.J. Zhang, G. Brinkschröder, and A. Knoll. Visuelles Reinforcement-Lernen zur Feinpositionierung eines Roboterarms über kompakte Zustandskodierung. In Tagungsband Autonome Mobile Robotersysteme, München, 1999.Google Scholar
- 6.M.-C. Nguyen and V. Graefe. Self-learning vision-guided robots for searching and grasping objects. In Proceedings of the IEEE International Conference on Robotics and Automation, pages 1633–1638, 2000.Google Scholar
- 9.Ch. J. C. H. Watkins. Learning from Delayed Rewards. PhD thesis, King’s College, Cambridge, England, 1989.Google Scholar
- 10.R. S. Sutton. Learning to predict by the method of temporal differences. Machine Learning, 3: 9–44, 1988.Google Scholar
- 12.R. F. Riesenfeld. Applications of B-Spline approximation to geometric problems of computer-aided design. PhD thesis, Syracuse University, 1973.Google Scholar
- 13.W. J. Gordon and R. F. Riesenfeld. Bspline curves and surfaces. In R. E. Barnhill and R. F. Riesenfeld, editors, Computer Aided Geometric Design. Academic Press, 1974.Google Scholar
- 14.M. Brown and C. J. Harris. Neural networks for modelling and control, chapter I, pages 17–55. In “Advances in Intelligent Control” , Ed., C. J. Harris, Taylor & Francis, London, 1994.Google Scholar