Heuristic dynamic programming for mobile robot path planning based on Dyna approach
Authors:
- 3 November 2016
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 3723-3730
Abstract
This paper presents a direct heuristic dynamic programming (HDP) based on Dyna planning (Dyna_HDP) for online model learning in a Markov decision process. This novel technique is composed of HDP policy learning to construct the Dyna agent for speeding up the learning time. We evaluate Dyna_HDP on a differential-drive wheeled mobile robot navigation problem in a 2D maze. The simulation is introduced to compare Dyna_HDP with other traditional reinforcement learning algorithms, namely one step Q-learning, Sarsa (λ), and Dyna_Q, under the same benchmark conditions. We demonstrate that Dyna_HDP has a faster near-optimal path than other algorithms, with high stability. In addition, we also confirm that the Dyna_HDP method can be applied in a multi-robot path planning problem. The virtual common environment model is learned from sharing the robots' experiences which significantly reduces the learning time.Keywords
This publication has 19 references indexed in Scilit:
- Data-driven heuristic dynamic programming with virtual realityNeurocomputing, 2015
- A three-network architecture for on-line learning and optimization based on adaptive dynamic programmingNeurocomputing, 2012
- Improving the Exploration Strategy in Bandit Algorithms2008
- A study of reinforcement learning with knowledge sharing for distributed autonomous system2004
- Dual heuristic programming excitation neurocontrol for generators in a multimachine power systemIEEE Transactions on Industry Applications, 2003
- Online learning control by association and reinforcementIEEE Transactions on Neural Networks, 2001
- Reinforcement Learning: An IntroductionIEEE Transactions on Neural Networks, 1998
- k-Certainty Exploration Method: an action selector to identify the environment in reinforcement learningArtificial Intelligence, 1997
- Adaptive critic designs: A case study for neurocontrolNeural Networks, 1995
- Integrated Architectures for Learning, Planning, and Reacting Based on Approximating Dynamic Programming1990