Multistrategy Learning for Robot Behaviours
作者： Claude Sammut , Tak Fai Yik
DOI: 10.1007/978-3-642-05177-7_23
关键词: Reinforcement learning 、 Robot 、 Machine learning 、 Planner 、 Humanoid robot 、 Constraint satisfaction problem 、 Robot learning 、 Artificial intelligence 、 Degrees of freedom 、 Learning classifier system 、 Computer science
摘要: Pure reinforcement learning does not scale well to domains with many degrees of freedom and particularly continuous domains. In this paper, we introduce a hybrid method in which symbolic planner constructs an approximate solution control problem. Subsequently, numerical optimisation algorithm is used refine the qualitative plan into operational policy. The demonstrated on problem stable walking gait for bipedal robot. We use approach illustrate benefits multistrategy robot learning.
springer.com
sci-hub.st 参考文章(34)
Min Sub Kim, William Uther, Automatic Gait Optimisation for Quadruped Robots ,(2003)
O. Hanagata, M. Fujita, S. Takamura, G. S. Hornby, T. Yamamoto, Autonomous evolution of gaits with the Sony Quadruped Robot genetic and evolutionary computation conference. pp. 1297- 1304 ,(1999)
Pat Langley, Dongkyu Choi, A unified cognitive architecture for physical agents national conference on artificial intelligence. pp. 1469- 1474 ,(2006)
Masaki Ogino, Koh Hosoda, Minoru Asada, Learning Energy-Efficient Walking with Ballistic Walking Adaptive Motion of Animals and Machines. pp. 155- 164 ,(2006) , 10.1007/4-431-31381-8_14
Thomas G. Dietterich, The MAXQ Method for Hierarchical Reinforcement Learning international conference on machine learning. pp. 118- 126 ,(1998)
Nils J. Nilsson, Scott Benson, Reacting, planning, and learning in an autonomous agent Machine intelligence. pp. 29- 62 ,(1996)
Ryszard S. Michalski, Inferential Theory of Learning: Developing Foundations for Multistrategy Learning ,(1992)
Claude Sammut, Bernhard Hengst, The Evolution of a Robot Soccer Team ISRR. pp. 517- 530 ,(2003) , 10.1007/3-540-36460-9_34
Lise Getoor, Ben Taskar, Introduction to statistical relational learning MIT Press. ,(2007)
Bernhard Hengst, Discovering Hierarchy in Reinforcement Learning with HEXQ international conference on machine learning. pp. 243- 250 ,(2002)