CASOG: Conservative Actor&#x2013;Critic With SmOoth Gradient for Skill Learning in Robot-Assisted Intervention

doi:10.1109/TIE.2023.3310021

CORC > 自动化研究所 > 中国科学院自动化研究所 > 多模态人工智能系统全国重点实验室

	CASOG: Conservative Actor–Critic With SmOoth Gradient for Skill Learning in Robot-Assisted Intervention
	Li, Hao 1,2; Zhou, Xiao-Hu 1,2; Xie, Xiao-Liang 1,2; Liu, Shi-Qi 1,2; Feng, Zhen-Qiu 1,2; Hou, Zeng-Guang 1,3
刊名	IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
	2023-09-18
页码	10
关键词	Deep neural network offline reinforcement learning robot-assisted intervention vascular robotic system
ISSN号	0278-0046
DOI	10.1109/TIE.2023.3310021
通讯作者	Zhou, Xiao-Hu(xiaohu.zhou@ia.ac.cn) ; Hou, Zeng-Guang(zengguang.hou@ia.ac.cn)
英文摘要	The robot-assisted intervention has shown reduced radiation exposure to physicians and improved precision in clinical trials. However, existing vascular robotic systems follow master-slave control mode and entirely rely on manual commands. This article proposes a novel offline reinforcement learning algorithm, Conservative Actor-critic with SmOoth Gradient (CASOG), to learn manipulation skills on vascular robotic systems. The proposed algorithm conservatively estimates Q-function and smooths gradients of convolution layers to deal with distribution shift and overfitting issues. Furthermore, to focus on complex manipulations, transitions with larger absolute temporal-difference error are sampled with higher probability. Comparative experiments on multiple vascular models and offline data demonstrate that CASOG delivers guidewire to the target with higher success rates and fewer backward steps than prior offline reinforcement learning methods. These results indicate that the proposed algorithm is promising to improve the autonomy of vascular robotic systems.
资助项目	National Natural Science Foundation of China
WOS研究方向	Automation & Control Systems ; Engineering ; Instruments & Instrumentation
语种	英语
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
WOS记录号	WOS:001122255700001
资助机构	National Natural Science Foundation of China
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/55030]
专题	多模态人工智能系统全国重点实验室
通讯作者	Zhou, Xiao-Hu; Hou, Zeng-Guang
作者单位	1.Chinese Acad Sci, Inst Automat, State Key Lab Multimodal Artificial Intelligence S, Beijing 100190, Peoples R China 2.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100049, Peoples R China 3.Macau Univ Sci & Technol, Inst Syst Engn, CASIA MUST Joint Lab Intelligence Sci & Technol, Macau, Peoples R China
推荐引用方式 GB/T 7714	Li, Hao,Zhou, Xiao-Hu,Xie, Xiao-Liang,et al. CASOG: Conservative Actor–Critic With SmOoth Gradient for Skill Learning in Robot-Assisted Intervention[J]. IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS,2023:10.
APA	Li, Hao,Zhou, Xiao-Hu,Xie, Xiao-Liang,Liu, Shi-Qi,Feng, Zhen-Qiu,&Hou, Zeng-Guang.(2023).CASOG: Conservative Actor–Critic With SmOoth Gradient for Skill Learning in Robot-Assisted Intervention.IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS,10.
MLA	Li, Hao,et al."CASOG: Conservative Actor–Critic With SmOoth Gradient for Skill Learning in Robot-Assisted Intervention".IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS (2023):10.