基于时空图神经网络的异构交通参与者风险预测

doi:10.19562/j.chinasae.qcgc.2024.09.001

摘要/Abstract

摘要：

有效预测驾驶员视野下的多交通参与者未来风险指标是为人类驾驶员提供风险预警，规避潜在碰撞风险的关键。大多数现有对风险的研究仅考虑场景中单一个体与本车之间的成对交互关系，并从评估而非预测的角度展开研究，而忽略异构交通参与者之间不同的交互关系及未来风险状态。本文提出了一种基于时空图卷积神经网络的异构多目标风险预测方法Risk-STGCN，通过图卷积及时间卷积分别对单帧场景图信息与时序信息进行学习，结合多层时序预测网络对多目标风险指标TTC进行预测。在开源BLVD与实车自采数据集上进行了训练验证，并与常用序列预测模型进行对比。实验结果表明，所提模型在不同数据集上的平均TTC误差均在0.95 s以下，多实验指标均优于文中所提到的其他模型，具有良好的鲁棒性，同时提升了复杂交通场景下风险预测的可解释性。

关键词: 智能汽车, 多交通参与者, 交互表征, 风险预测, 时空图神经网络

Abstract:

Effectively predicting the future risk indicators of multiple traffic participants under the driver's field of vision is the key to providing risk warnings to human drivers and avoiding potential collision risk. Most existing research on risk only considers the pairwise interaction between a single individual and the vehicle in the scene， and conducts research from the perspective of evaluation rather than prediction， while ignoring the different interaction between heterogeneous traffic participants and future risk status. This paper proposes a heterogeneous multi-objective risk prediction method Risk-STGCN based on spatiotemporal graph convolutional neural network， using graph convolution and temporal convolution to learn single-frame scene graph information and timing information respectively， combined with multi-layer timing prediction network to predict the multi-objective risk indicator TTC. Training and verification are conducted on the open source data set BLVD and the real vehicle self-collected data set， which is then compared with commonly used sequence prediction models. The experimental results show that the average TTC error of the proposed model on different data sets is less than 0.95 s， with multiple experimental indicators better than other models mentioned in this paper. The proposed model has good robustness and improves the interpretability of risk prediction in complex traffic scenarios.

Key words: intelligent vehicles, multiple traffic participants, interactive representation, risk prediction, spatio-temporal graph neural network

孟相浩,牛凌,席军强,陈丹妮,吕超. 基于时空图神经网络的异构交通参与者风险预测[J]. 汽车工程, 2024, 46(9): 1537-1545.

Xianghao Meng,Ling Niu,Junqiang Xi,Danni Chen,Chao Lü. Risk Prediction of Heterogeneous Traffic Participants Based on Spatio-Temporal Graph Neural Networks[J]. Automotive Engineering, 2024, 46(9): 1537-1545.

图/表 12

图1

图2

图3

图4

图5

表1

超参数设置"

优化器	激活函数	学习率	Batch大小	Epoch	$α$	GCN层数	TSP层数
SGD	PReLU	0.01/0.002	128	250	1	1	5

表1

图6

图7

图8

图9

表2

图10

参考文献 30

1	PUGACHEV I， KULIKOV Y， CHEGLOV V. Features of traffic organization and traffic safety in cities［J］. Transportation Research Procedia， 2020， 50： 766-772.
2	GUO J， KURUP U， SHAH M. Is it safe to drive？ an overview of factors， metrics， and datasets for driveability assessment in autonomous driving［J］. IEEE Transactions on Intelligent Transportation Systems， 2019， 21（8）： 3135-3151.
3	BUCSUHÁZY K， MATUCHOVÁ E， ZŮVALA R， et al. Human factors contributing to the road traffic accident occurrence［J］. Transportation Research Procedia， 2020， 45： 555-561.
4	KATRAKAZAS C， QUDDUS M， CHEN W H. A new integrated collision risk assessment methodology for autonomous vehicles［J］. Accident Analysis & Prevention， 2019， 127： 61-79.
5	YU M Y， VASUDEVAN R， JOHNSON-ROBERSON M. Risk assessment and planning with bidirectional reachability for autonomous driving［C］. 2020 IEEE International Conference on Robotics and Automation （ICRA）， 2020： 5363-5369.
6	RYAN C， MURPHY F， MULLINS M. End-to-end autonomous driving risk analysis： a behavioural anomaly detection approach［J］. IEEE Transactions on Intelligent Transportation Systems， 2020， 22（3）： 1650-1662.
7	OLMUŞ H， ERBAŞ S. Analysis of traffic accidents caused by drivers by using Log-linear models［J］. Promet-Traffic&Transportation， 2012， 24（6）： 495-504.
8	QIAO C， POLLMEYER S， WANG Y， et al. Model predictive trajectory planning of autonomous vehicles considering dynamic driving constraints［C］. 2020 4th CAA International Conference on Vehicular Control and Intelligence （CVCI）， 2020： 136-141.
9	MIGLANI A， KUMAR N. Deep learning models for traffic flow prediction in autonomous vehicles： a review， solutions， and challenges［J］. Vehicular Communications， 2019， 20： 100184.
10	JOERER S， SEGATA M， BLOESSL B， et al. A vehicular networking perspective on estimating vehicle collision probability at intersections［J］. IEEE Transactions on Vehicular Technology， 2013， 63（4）： 1802-1812.
11	CHEN C， LÜ N， LIU L， et al. Critical safe distance design to improve driving safety based on vehicle-to-vehicle communications［J］. Journal of Central South University， 2013， 20（11）： 3334-3344.
12	张哲雨，吕超，李景行，等. 基于车辆视角数据的行人轨迹预测与风险等级评定［J］. 汽车工程， 2022， 44（5）： 675-683.
	ZHANG Zheyu，LÜ Chao， LI Jinghang， et al. Pedestrian trajectory prediction and risk grade assessment based on vehicle-perspective pedestrian data［J］. Automotive Engineering， 2022， 44（5）： 675-683.
13	KUSANO K D， GABLER H C. Safety benefits of forward collision warning， brake assist， and autonomous braking systems in rear-end collisions［J］. IEEE Transactions on Intelligent Transportation Systems， 2012， 13（4）： 1546-1555.
14	BOARD N. Special investigation report-highway vehicle and infrastructure-based technology for the prevention of rear-end collisions［J］. NTSB Number SIR-OI/ll， 2001.
15	李霖，朱西产，马志雄. 驾驶员在真实交通危险工况中的制动反应时间［J］. 汽车工程， 2014， 36（10）： 1225-1229.
	LI Lin， ZHU Xichan， MA Zhixiong. Driver brake reaction time under real traffic risk scenarios［J］. Automotive Engineering， 2014， 36（10）： 1225-1229.
16	LEFÈVRE S， VASQUEZ D， LAUGIER C. A survey on motion prediction and risk assessment for intelligent vehicles［J］. ROBOMECH Journal， 2014， 1（1）： 1-14.
17	GUO M， ZHAO X， YAO Y， et al. A study of freeway crash risk prediction and interpretation based on risky driving behavior and traffic flow data［J］. Accident Analysis & Prevention， 2021， 160： 106328.
18	ZHANG L， YUAN K， CHU H， et al. Pedestrian collision risk assessment based on state estimation and motion prediction［J］. IEEE Transactions on Vehicular Technology， 2021， 71（1）： 98-111.
19	SHEN X， RAKSINCHAROENSAK P. Pedestrian-aware statistical risk assessment［J］. IEEE Transactions on Intelligent Transportation Systems， 2021， 23（7）： 7910-7918.
20	LI Z， GONG J， LU C， et al. Interactive behavior prediction for heterogeneous traffic participants in the urban road： a graph-neural-network-based multitask learning framework［J］. IEEE/ASME Transactions on Mechatronics， 2021， 26（3）： 1339-1349.
21	RUMMELHARD L， NÈGRE A， PERROLLAZ M， et al. Probabilistic grid-based collision risk prediction for driving application［C］. Experimental Robotics： The 14th International Symposium on Experimental Robotics， 2016： 821-834.
22	ROTH M， FLOHR F， GAVRILA D M. Driver and pedestrian awareness-based collision risk analysis［C］. 2016 IEEE Intelligent Vehicles Symposium （IV）， 2016： 454-459.
23	王建强，吴剑，李洋. 基于人-车-路协同的行车风险场概念，原理及建模［J］. 中国公路学报， 2016， 29（1）： 105-114.
	WANG Jianqiang， WU Jian， LI Yang. Concept， principle and modeling of driving risk field based on driver-vehicle-road interaction［J］. China Journal of Highway and Transport， 2016， 29（1）： 105-114.
24	JASOUR A， HUANG X， WANG A， et al. Fast nonlinear risk assessment for autonomous vehicles using learned conditional probabilistic models of agent futures［J］. Autonomous Robots， 2022： 1-14.
25	MOHAMED A， QIAN K， ELHOSEINY M， et al. Social-stgcnn： a social spatio-temporal graph convolutional neural network for human trajectory prediction［C］. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition， 2020： 14424-14432.
26	KOSARAJU V， SADEGHIAN A， MARTÍN-MARTÍN R， et al. Social-bigat： multimodal trajectory forecasting using bicycle-gan and graph attention networks［J］. Advances in Neural Information Processing Systems， 2019， 32.
27	ANAYA J J， MERDRIGNAC P， SHAGDAR O， et al. Vehicle to pedestrian communications for protection of vulnerable road users［C］. 2014 IEEE Intelligent Vehicles Symposium Proceedings， 2014： 1037-1042.
28	XUE J， FANG J， LI T， et al. BLVD： building a large-scale 5D semantics benchmark for autonomous driving［C］. 2019 International Conference on Robotics and Automation （ICRA）， 2019： 6685-6691.
29	ZHANG H， NAN Z， YANG T， et al. A driving behavior recognition model with bi-LSTM and multi-scale CNN［C］. 2020 IEEE Intelligent Vehicles Symposium （IV）， 2020： 284-289.
30	YANG T， NAN Z， ZHANG H， et al. Traffic agent trajectory prediction using social convolution and attention mechanism ［J］. arXiv preprint arXiv：.

场景	车辆		行人
场景	ATE/s	FTE/s	ATE/s	FTE/s
BLVD直行	0.982	1.274	0.690	0.934
实车左转	1.376	2.032	0.832	1.712
实车直行	0.755	1.247	0.671	1.017
实车右转	1.071	1.672	0.791	1.434

[1]	赵静,梁浩,徐天啸,肖雅月,姜博文. 面向智能汽车赛博物理系统的领域特定建模语言研究[J]. 汽车工程, 2024, 46(8): 1370-1381.
[2]	朱甲梁,刘巧斌,杨帆,杨路,李巍华. 异质车型影响下智能汽车二维碰撞风险预测[J]. 汽车工程, 2024, 46(8): 1414-1421.
[3]	宋东鉴,赵健,朱冰,佟静,韩嘉懿,刘斌. 考虑风险时空分布特征的跟驰工况行车风险预测[J]. 汽车工程, 2024, 46(5): 766-775.
[4]	朱冰,黄殷梓,赵健,张培兴,薛经纬. 面向车载相机采集图像的智能汽车测试场景关键性量化模型[J]. 汽车工程, 2024, 46(4): 557-563.
[5]	丁志杰,王亚飞,章翼辰,邬明宇,王亦乐. 基于复合动态采样的自动驾驶矿车节能路径规划方法[J]. 汽车工程, 2024, 46(4): 588-595.
[6]	郝建平,苏炎召,钟志华,黄晋. 面向智能汽车的SOA架构及服务调度机制研究[J]. 汽车工程, 2023, 45(9): 1563-1572.
[7]	赵高士,陈龙,蔡英凤,廉玉波,王海,刘擎超,滕成龙. 融合复杂网络和记忆增强网络的轨迹预测技术[J]. 汽车工程, 2023, 45(9): 1608-1616.
[8]	胡启慧,蔡英凤,王海,陈龙,董钊志,刘擎超. 基于层次图注意的异构多目标轨迹预测方法[J]. 汽车工程, 2023, 45(8): 1448-1456.
[9]	高翔,陈龙,王歆叶,熊晓夏,李祎承,陈月霞. 基于轨迹预测的智能汽车行驶风险评估方法[J]. 汽车工程, 2023, 45(4): 588-597.
[10]	胡杰,朱琪,陈锐鹏,张敏超,张志豪,刘昊岩. 引入必经点约束的智能汽车全局路径规划[J]. 汽车工程, 2023, 45(3): 350-360.
[11]	张紫微,郑玲,李以农,乔旭强,郑浩,王戡. 考虑前车运动不确定性的多目标自适应巡航控制[J]. 汽车工程, 2023, 45(3): 361-371.
[12]	刘正发,吴亚,刘佩根,顾荣琦,陈广. 基于特征和标签联合分布匹配的智能驾驶跨域自适应目标检测[J]. 汽车工程, 2023, 45(11): 2082-2091.
[13]	赵树廉,来飞,李克强,陈涛,孟璋劼,唐逸超,吴思宇,田浩东. 基于数字孪生技术的智能汽车测试方法研究[J]. 汽车工程, 2023, 45(1): 42-51.
[14]	邵文博,李骏,张玉新,王红. 智能汽车预期功能安全保障关键技术[J]. 汽车工程, 2022, 44(9): 1289-1304.
[15]	汪梓豪,蔡英凤,王海,陈龙,熊晓夏. 基于单目视觉运动估计的周边多目标轨迹预测方法[J]. 汽车工程, 2022, 44(9): 1318-1326.