CSR: Cascade Conditional Variational Auto Encoder with Socially-aware Regression for Pedestrian Trajectory Prediction

doi:10.1016/j.patcog.2022.109030

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 > 机器人应用与理论组

	CSR: Cascade Conditional Variational Auto Encoder with Socially-aware Regression for Pedestrian Trajectory Prediction
	Zhou, Hao 2,4; Ren, Dongchun 3; Yang, Xu 4; Fan, Mingyu 1,3; Huang, Hai 2
刊名	PATTERN RECOGNITION
	2023
卷号	133 页码:10
关键词	Pedestrian trajectory prediction Socially -aware model Conditional variational autoencoder (CVAE)
ISSN号	0031-3203
DOI	10.1016/j.patcog.2022.109030
通讯作者	Fan, Mingyu(fanmingyu@wzu.edu.cn) ; Huang, Hai(haihus@163.com)
英文摘要	Pedestrian trajectory prediction is a key technology in many real applications such as video surveillance, social robot navigation, and autonomous driving, and significant progress has been made in this research topic. However, there remain two limitations of previous studies. First, the losses of the last time steps are heavier weighted than that of the beginning time steps in the objective function at the learning stage, causing the prediction errors generated at the beginning to accumulate to large errors at the last time steps at the inference stage. Second, the prediction results of multiple pedestrians in the prediction horizon might be socially incompatible with the interactions modeled by past trajectories. To overcome these limitations, this work proposes a novel trajectory prediction method called CSR, which consists of a cascaded conditional variational autoencoder (CVAE) module and a socially-aware regression module. The CVAE module estimates the future trajectories in a cascaded sequential manner. Specifically, each CVAE concatenates the past trajectories and the predicted location points so far as the input and predicts the adjacent location at the following time step. The socially-aware regression module generates offsets from the estimated future trajectories to produce the corrected predictions, which are more reasonable and accurate than the estimated trajectories. Experiments results demonstrate that the proposed method exhibits significant improvements over state-of-the-art methods on the Stanford Drone Dataset (SDD) and the ETH/UCY dataset of approximately 38.0% and 22.2%, respectively. The code is available at https: //github.com/zhouhao94/CSR . (c) 2022 Elsevier Ltd. All rights reserved.
资助项目	National Natural Science Foundation of China[U21A20490] ; National Natural Science Foundation of China[61633009] ; National Natural Science Foundation of China[61973301] ; National Natural Science Foundation of China[61972020] ; National Natural Science Foundation of China[61772373] ; National Natural Science Foundation of China[51579053] ; National Natural Science Foundation of China[U1613213] ; Beijing Nova Program[Z20110 0 0 06820 046] ; Meituan Open RD Fund
WOS研究方向	Computer Science ; Engineering
语种	英语
出版者	ELSEVIER SCI LTD
WOS记录号	WOS:000861386400001
资助机构	National Natural Science Foundation of China ; Beijing Nova Program ; Meituan Open RD Fund
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/50406]
专题	自动化研究所_复杂系统管理与控制国家重点实验室_机器人应用与理论组
通讯作者	Fan, Mingyu; Huang, Hai
作者单位	1.Wenzhou Univ, Coll Comp Sci, Wenzhou, Peoples R China 2.Harbin Engn Univ, Natl Key Lab Sci & Technol Underwater Vehicle, Harbin, Peoples R China 3.Meituan, Res Ctr Autonomous Vehicles, Beijing, Peoples R China 4.Chinese Acad Sci, Inst Automat, State Key Lab Management & Control Complex Syst, Beijing, Peoples R China
推荐引用方式 GB/T 7714	Zhou, Hao,Ren, Dongchun,Yang, Xu,et al. CSR: Cascade Conditional Variational Auto Encoder with Socially-aware Regression for Pedestrian Trajectory Prediction[J]. PATTERN RECOGNITION,2023,133:10.
APA	Zhou, Hao,Ren, Dongchun,Yang, Xu,Fan, Mingyu,&Huang, Hai.(2023).CSR: Cascade Conditional Variational Auto Encoder with Socially-aware Regression for Pedestrian Trajectory Prediction.PATTERN RECOGNITION,133,10.
MLA	Zhou, Hao,et al."CSR: Cascade Conditional Variational Auto Encoder with Socially-aware Regression for Pedestrian Trajectory Prediction".PATTERN RECOGNITION 133(2023):10.