ABSTRACT
XCS has been shown to be an effective genetics-based classification, datamining, and reinforcement learning tool. The systems learns suitable, compact, maximally general problem solutions online. In the robotics and cognitive systems domains, however, applications of XCSF are very sparse and mostly restricted to small, symbolic problems. Recently, a sensorimotor XCSF system was applied to cognitive arm control. In this paper, we show how this XCSF-based armcontrol mechanisms can be extended (1) to efficiently exploit redundant behavioral alternatives and (2) to guide the control of dynamic arm plants. The XCSF system encodes redundant alternatives in its inverse control representations and resolves the encoded redundancies dependent on current constraints--such as arm posture preferences - on the fly. An adaptive PD controller translates the XCSF-based direction and distance commands into actual motor commands for dynamic arm control. We apply the complete system to the control of a simulated, physical arm with three degrees of freedom in a two-dimensional environment and to a simulation of the industrial KR16 Kuka arm with ODE-based physics engine.
- K. J. Åström and B. Wittenmark. Adaptive Control. Addison-Wesley, second edition, 1995.Google Scholar
- M. V. Butz and O. Herbort. Context-dependent predictions and cognitive arm control with XCSF. Genetic and Evolutionary Computation Conference, GECCO 2008, pages 1357--1364, 2008. Google ScholarDigital Library
- M. V. Butz, O. Herbort, and J. Hoffmann. Exploiting redundancy for flexible behavior: Unsupervised learning in a modular sensorimotor control architecture. Psychological Review, 114:1015--1046, 2007.Google ScholarCross Ref
- M. V. Butz, P. L. Lanzi, and S. W. Wilson. Function approximation with XCS: Hyperellipsoidal conditions, recursive least squares, and compaction. IEEE Transactions on Evolutionary Computation, 12:355--376, 2008. Google ScholarDigital Library
- G. F. Franklin, J. D. Powell, and A. Emami-Naeini. Feedback Control of Dynamic Systems. Prentice Hall, Upper Saddle River, NJ, 4th edition, 2002. Google ScholarDigital Library
- M. S. A. Graziano. The organization of behavioral repertoire in motor cortex. Annual Review of Neuroscience, 29:105--134, 2006.Google ScholarCross Ref
- R. Kelly. PD control with desired gravity compensation of robotic manipulators: A review. The International Journal of Robotics Research, 16:660--672, 1997. Google ScholarDigital Library
- T. Kovacs. XCS classifier system reliably evolves accurate, complete, and minimal representations for boolean functions. In Roy, Chawdhry, and Pant, editors, Soft computing in engineering design and manufacturing, pages 59--68. Springer-Verlag, Berlin Heidelberg, 1997.Google Scholar
- KUKA. Kuka kr16, 2008. KUKA Roboter GmbH, http://www.kuka.com/germany/de/products/newline industrial_robots/low/kr16/.Google Scholar
- P. L. Lanzi, D. Loiacono, S. W. Wilson, and D. E. Goldberg. Generalization in the XCSF classifier system: Analysis, improvement, and extension. Evolutionary Computation, 15:133--168, 2007. Google ScholarDigital Library
- P. L. Lanzi and S. W. Wilson. Using convex hulls to represent classifier conditions. Genetic and Evolutionary Computation Conference, GECCO 2006, pages 1481--1488, 2006. Google ScholarDigital Library
- C. Ltd. Webots reference manual, 2008. www.cyberbotics.com.Google Scholar
- A. Orriols-Puig and E. Bernadó-Mansilla. Bounding XCS's parameters for unbalanced datasets. Genetic and Evolutionary Computation Conference, GECCO 2006, pages 1561--1568, 2006. Google ScholarDigital Library
- P. Stalph and M. V. Butz. Documentation of XCSF-Ellipsoids Java plus visualization. MEDAL Report 2008008, Missouri Estimation of Distribution Algorithms Laboratory, University of Missouri in St. Louis, MO, 2008.Google Scholar
- S. W. Wilson. Classifier fitness based on accuracy. Evolutionary Computation, 3(2):149--175, 1995. Google ScholarDigital Library
- S. W. Wilson. Generalization in the XCS classifier system. Genetic Programming 1998: Proceedings of the Third Annual Conference, pages 665--674, 1998.Google Scholar
- S. W. Wilson. Classifiers that approximate functions. Natural Computing, 1:211--234, 2002. Google ScholarDigital Library
Index Terms
- Learning sensorimotor control structures with XCSF: redundancy exploitation and dynamic control
Recommendations
Filtering sensory information with xcsf: Improving learning robustness and robot arm control performance
It has been shown previously that the control of a robot arm can be efficiently learned using the XCSF learning classifier system, which is a nonlinear regression system based on evolutionary computation. So far, however, the predictive knowledge about ...
On the scalability of XCS(F)
GECCO '09: Proceedings of the 11th Annual conference on Genetic and evolutionary computationMany successful applications have proven the potential of Learning Classifier Systems and the XCS classifier system in particular in datamining, reinforcement learning, and function approximation tasks. Recent research has shown that XCS is a highly ...
Controlling a four degree of freedom arm in 3D using the XCSF learning classifier system
KI'09: Proceedings of the 32nd annual German conference on Advances in artificial intelligenceThis paper shows for the first time that a Learning Classifier System, namely XCSF, can learn to control a realistic arm model with four degrees of freedom in a three-dimensional workspace. XCSF learns a locally linear approximation of the Jacobian of ...
Comments