Reinforcement Learning with Dynamic Movement Primitives for Obstacle Avoidance

doi:10.3390/app112311184

CIOMP OpenIR

	Reinforcement Learning with Dynamic Movement Primitives for Obstacle Avoidance
	A. Li; Z. Z. Liu; W. R. Wang; M. C. Zhu; Y. H. Li; Q. Huo and M. Dai
	2021
发表期刊	Applied Sciences-Basel
卷号	11 期号:23 页码:13
摘要	Dynamic movement primitives (DMPs) are a robust framework for movement generation from demonstrations. This framework can be extended by adding a perturbing term to achieve obstacle avoidance without sacrificing stability. The additional term is usually constructed based on potential functions. Although different potentials are adopted to improve the performance of obstacle avoidance, the profiles of potentials are rarely incorporated into reinforcement learning (RL) framework. In this contribution, we present a RL based method to learn not only the profiles of potentials but also the shape parameters of a motion. The algorithm employed is PI2 (Policy Improvement with Path Integrals), a model-free, sampling-based learning method. By using the PI2, the profiles of potentials and the parameters of the DMPs are learned simultaneously; therefore, we can optimize obstacle avoidance while completing specified tasks. We validate the presented method in simulations and with a redundant robot arm in experiments.
DOI	10.3390/app112311184
URL	查看原文
收录类别	sci
语种	英语
引用统计
文献类型	期刊论文
条目标识符	http://ir.ciomp.ac.cn/handle/181722/67031
专题	中国科学院长春光学精密机械与物理研究所
推荐引用方式 GB/T 7714	A. Li,Z. Z. Liu,W. R. Wang,et al. Reinforcement Learning with Dynamic Movement Primitives for Obstacle Avoidance[J]. Applied Sciences-Basel,2021,11(23):13.
APA	A. Li,Z. Z. Liu,W. R. Wang,M. C. Zhu,Y. H. Li,&Q. Huo and M. Dai.(2021).Reinforcement Learning with Dynamic Movement Primitives for Obstacle Avoidance.Applied Sciences-Basel,11(23),13.
MLA	A. Li,et al."Reinforcement Learning with Dynamic Movement Primitives for Obstacle Avoidance".Applied Sciences-Basel 11.23(2021):13.

条目包含的文件		下载所有文件
文件名称/大小	文献类型	版本类型	开放类型	使用许可
Reinforcement Learni（1107KB）	期刊论文	出版稿	开放获取	CC BY-NC-SA	浏览下载