Modular Reinforcement Learning: A Case Study in a Robot Domain

T h e behaviour of reinforcement learning (RL) algorithms is best understood in completely observable, finite stateand action-space, discrete-time controlled Markov-chains. Robot-learning domains, on the other hand, are inherently infinite both in time and space, and moreover they are only partially observable. In this article we suggest a systematic design method whose motivation comes from the desire to transform the task-to-be-solved into a finite-state, discrete-time, "approximately" Markovian task, which is completely observable, too. The key idea is to break up the problem into subtasks and design controllers for each of the subtasks. Then operating conditions are attached to the controllers (together the controllers and their operating conditions which are called modules) and possible additional features are designed to facilitate observability. A new discrete time-counter is introduced at the "module-level" that clicks only when a change in the value of one of the features is observed. The approach was tried out on a real-life robot. Several RL algorithms were compared and it was found that a model-based approach worked best. The learnt switching strategy performed equally well as a handcrafted version. Moreover, the learnt strategy seemed to exploit certain properties of the environment which could not have been seen in advance, which predicted the promising possibility that a learnt controller might overperform a handcrafted switching strategy in the future.

[1]  Richard S. Sutton,et al.  Temporal credit assignment in reinforcement learning , 1984 .

[2]  Rodney A. Brooks,et al.  Elephants don't play chess , 1990, Robotics Auton. Syst..

[3]  Pattie Maes,et al.  A bottom-up mechanism for behavior selection in an artificial creature , 1991 .

[4]  Sridhar Mahadevan,et al.  Automatic Programming of Behavior-Based Robots Using Reinforcement Learning , 1991, Artif. Intell..

[5]  D. Sofge THE ROLE OF EXPLORATION IN LEARNING CONTROL , 1992 .

[6]  András Lörincz,et al.  Behavior of an Adaptive Self-organizing Autonomous Agent Working with Cues and Competing Concepts , 1993, Adapt. Behav..

[7]  Michael I. Jordan,et al.  MASSACHUSETTS INSTITUTE OF TECHNOLOGY ARTIFICIAL INTELLIGENCE LABORATORY and CENTER FOR BIOLOGICAL AND COMPUTATIONAL LEARNING DEPARTMENT OF BRAIN AND COGNITIVE SCIENCES , 1996 .

[8]  Z. Kalmar,et al.  Generalization in an autonomous agent , 1994, Proceedings of 1994 IEEE International Conference on Neural Networks (ICNN'94).

[9]  Michael I. Jordan,et al.  Learning Without State-Estimation in Partially Observable Markovian Decision Processes , 1994, ICML.

[10]  Leslie Pack Kaelbling,et al.  Learning Policies for Partially Observable Environments: Scaling Up , 1997, ICML.

[11]  Richard S. Sutton,et al.  Generalization in Reinforcement Learning: Successful Examples Using Sparse Coarse Coding , 1995, NIPS.

[12]  Andrew G. Barto,et al.  Learning to Act Using Real-Time Dynamic Programming , 1995, Artif. Intell..

[13]  Csaba Szepesvári,et al.  A Generalized Reinforcement-Learning Model: Convergence and Applications , 1996, ICML.

[14]  Michael L. Littman,et al.  Algorithms for Sequential Decision Making , 1996 .

[15]  John N. Tsitsiklis,et al.  Analysis of Temporal-Diffference Learning with Function Approximation , 1996, NIPS.

[16]  Minoru Asada,et al.  Behavior coordination for a mobile robot using modular reinforcement learning , 1996, Proceedings of IEEE/RSJ International Conference on Intelligent Robots and Systems. IROS '96.

[17]  Maja J. Mataric,et al.  Reinforcement Learning in the Multi-Robot Domain , 1997, Auton. Robots.

[18]  Csaba Szepesvári,et al.  A Unified Analysis of Value-Function-Based Reinforcement-Learning Algorithms , 1999, Neural Computation.