Safe Motion Planning for Autonomous Vehicles by Quantifying Uncertainties of Deep Learning-Enabled Environment Perception

doi:10.1109/TIV.2023.3297735

CORC > 自动化研究所 > 中国科学院自动化研究所 > 复杂系统管理与控制国家重点实验室 > 先进控制与自动化团队

	Safe Motion Planning for Autonomous Vehicles by Quantifying Uncertainties of Deep Learning-Enabled Environment Perception
	Li, Dachuan 5; Liu, Bowen 4; Huang, Zijian 4; Hao, Qi 4,5; Zhao, Dezong 3; Tian, Bin 1,2
刊名	IEEE TRANSACTIONS ON INTELLIGENT VEHICLES
	2024
卷号	9 期号:1 页码:2318-2332
关键词	Uncertainty Planning Three-dimensional displays Decision making Detectors Trajectory Probabilistic logic Autonomous driving motion planning object detection Bayesian deep learning uncertainty quantification
ISSN号	2379-8858
DOI	10.1109/TIV.2023.3297735
通讯作者	Hao, Qi(haoq@sustech.edu.cn)
英文摘要	Conventional perception-planning pipelines of autonomous vehicles (AV) utilize deep learning (DL) techniques that typically generate deterministic outputs without explicitly evaluating their uncertainties and trustworthiness. Therefore, the downstream decision-making components may generate unsafe outputs leading to system failure or accidents, if the preceding perception component provides highly uncertain information. To mitigate this issue, this article proposes a coherent safe perception-planning framework that quantifies and transfers DL-based perception uncertainties. Following the Bayesian Deep Learning paradigm, we design a probabilistic 3D object detector that extracts objects from LiDAR point clouds while quantifying the corresponding aleatoric and epistemic uncertainty. A chance-constrained motion planner is designed to formulate an explicit link between DL-based perception uncertainties and operation risk and generate safe and risk-bounding trajectories. The proposed framework is validated through various challenging scenarios in the CARLA simulator. Experiment results demonstrate that our framework can effectively capture the uncertainties in DL, and generate trajectories that bound the risk under DL perception uncertainties. It also outperforms counterpart approaches without explicitly evaluating the uncertainties of DL-based perception.
资助项目	National Natural Science Foundation of China[52272419] ; National Natural Science Foundation of China[62261160654] ; Science and Technology Innovation Committee of Shenzhen City[JCYJ20200109141622964] ; Science and Technology Innovation Committee of Shenzhen City[JCYJ20220818103006012] ; National Key Research and Development Program of China[2022YFB4703700]
WOS研究方向	Computer Science ; Engineering ; Transportation
语种	英语
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
WOS记录号	WOS:001173317800192
资助机构	National Natural Science Foundation of China ; Science and Technology Innovation Committee of Shenzhen City ; National Key Research and Development Program of China
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/58777]
专题	自动化研究所_复杂系统管理与控制国家重点实验室_先进控制与自动化团队
通讯作者	Hao, Qi
作者单位	1.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100190, Peoples R China 2.Chinese Acad Sci, Inst Automat, State Key Lab Management & Control Complex Syst, Beijing 100190, Peoples R China 3.Univ Glasgow, James Watt Sch Engn, Glasgow G12 8QQ, Lanark, Scotland 4.Southern Univ Sci & Technol, Dept Comp Sci & Engn, Shenzhen 518057, Peoples R China 5.Southern Univ Sci & Technol, Res Inst Trustworthy Autonomous Syst, Shenzhen 518055, Peoples R China
推荐引用方式 GB/T 7714	Li, Dachuan,Liu, Bowen,Huang, Zijian,et al. Safe Motion Planning for Autonomous Vehicles by Quantifying Uncertainties of Deep Learning-Enabled Environment Perception[J]. IEEE TRANSACTIONS ON INTELLIGENT VEHICLES,2024,9(1):2318-2332.
APA	Li, Dachuan,Liu, Bowen,Huang, Zijian,Hao, Qi,Zhao, Dezong,&Tian, Bin.(2024).Safe Motion Planning for Autonomous Vehicles by Quantifying Uncertainties of Deep Learning-Enabled Environment Perception.IEEE TRANSACTIONS ON INTELLIGENT VEHICLES,9(1),2318-2332.
MLA	Li, Dachuan,et al."Safe Motion Planning for Autonomous Vehicles by Quantifying Uncertainties of Deep Learning-Enabled Environment Perception".IEEE TRANSACTIONS ON INTELLIGENT VEHICLES 9.1(2024):2318-2332.