Magnetic Field-Based Reward Shaping for Goal-Conditioned Reinforcement Learning

doi:10.1109/JAS.2023.123477

CORC > 自动化研究所 > 中国科学院自动化研究所 > 学术期刊 > IEEE/CAA Journal of Automatica Sinica

	Magnetic Field-Based Reward Shaping for Goal-Conditioned Reinforcement Learning
	Hongyu Ding; Yuanze Tang; Qing Wu; Bo Wang; Chunlin Chen; Zhi Wang
刊名	IEEE/CAA Journal of Automatica Sinica
	2023
卷号	10 期号:12 页码:2233-2247
关键词	Dynamic environments goal-conditioned reinforcement learning magnetic field reward shaping
ISSN号	2329-9266
DOI	10.1109/JAS.2023.123477
英文摘要	Goal-conditioned reinforcement learning (RL) is an interesting extension of the traditional RL framework, where the dynamic environment and reward sparsity can cause conventional learning algorithms to fail. Reward shaping is a practical approach to improving sample efficiency by embedding human domain knowledge into the learning process. Existing reward shaping methods for goal-conditioned RL are typically built on distance metrics with a linear and isotropic distribution, which may fail to provide sufficient information about the ever-changing environment with high complexity. This paper proposes a novel magnetic field-based reward shaping (MFRS) method for goal-conditioned RL tasks with dynamic target and obstacles. Inspired by the physical properties of magnets, we consider the target and obstacles as permanent magnets and establish the reward function according to the intensity values of the magnetic field generated by these magnets. The nonlinear and anisotropic distribution of the magnetic field intensity can provide more accessible and conducive information about the optimization landscape, thus introducing a more sophisticated magnetic reward compared to the distance-based setting. Further, we transform our magnetic reward to the form of potential-based reward shaping by learning a secondary potential function concurrently to ensure the optimal policy invariance of our method. Experiments results in both simulated and real-world robotic manipulation tasks demonstrate that MFRS outperforms relevant existing methods and effectively improves the sample efficiency of RL algorithms in goal-conditioned tasks with various dynamics of the target and obstacles.
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/52467]
专题	自动化研究所_学术期刊_IEEE/CAA Journal of Automatica Sinica
推荐引用方式 GB/T 7714	Hongyu Ding,Yuanze Tang,Qing Wu,et al. Magnetic Field-Based Reward Shaping for Goal-Conditioned Reinforcement Learning[J]. IEEE/CAA Journal of Automatica Sinica,2023,10(12):2233-2247.
APA	Hongyu Ding,Yuanze Tang,Qing Wu,Bo Wang,Chunlin Chen,&Zhi Wang.(2023).Magnetic Field-Based Reward Shaping for Goal-Conditioned Reinforcement Learning.IEEE/CAA Journal of Automatica Sinica,10(12),2233-2247.
MLA	Hongyu Ding,et al."Magnetic Field-Based Reward Shaping for Goal-Conditioned Reinforcement Learning".IEEE/CAA Journal of Automatica Sinica 10.12(2023):2233-2247.