适用于多车交互场景的车辆轨迹预测模型

doi:10.13229/j.cnki.jdxbgxb.20220728

吉林大学学报(工学版) ›› 2024, Vol. 54 ›› Issue (5): 1188-1195.doi: 10.13229/j.cnki.jdxbgxb.20220728

• 车辆工程·机械工程 • 上一篇

适用于多车交互场景的车辆轨迹预测模型

黄玲^1,²(),崔躜¹,游峰^1,³(),洪佩鑫¹,钟浩川¹,曾译萱¹

^1.华南理工大学土木与交通学院，广州 510640
^2.东南大学现代城市交通技术江苏高校协同创新中心，南京 210096
^3.人工智能与数字经济广东省实验室，广州 510330

收稿日期:2022-06-10 出版日期:2024-05-01 发布日期:2024-06-11
通讯作者: 游峰 E-mail:hling@scut.edu.cn;youfeng@scut.edu.cn
作者简介:黄玲（1979-），女，副教授，博士.研究方向：自动驾驶，驾驶行为建模，交通仿真.E-mail： hling@scut.edu.cn
基金资助:
广东省基础与应用基础研究基金项目(2023A1515010742)

Vehicle trajectory prediction model for multi-vehicle interaction scenario

Ling HUANG^1,²(),Zuan CUI¹,Feng YOU^1,³(),Pei-xin HONG¹,Hao-chuan ZHONG¹,Yi-xuan ZENG¹

^1.School of Civil Engineering and Transportation，South China University of Technology，Guangzhou 510640，China
^2.Jiangsu Key Laboratory of Urban ITS，Southeast University，Nanjing 210096，China
^3.Pazhou Lab，Guangzhou 510330，China

Received:2022-06-10 Online:2024-05-01 Published:2024-06-11
Contact: Feng YOU E-mail:hling@scut.edu.cn;youfeng@scut.edu.cn

摘要/Abstract

摘要：

提出了一种具有动态交互感知池化层的多长短期记忆神经网络（DIP-LSTM）模型结构，使得场景中相邻的车辆通过池化（Pooling）共享各自LSTM网络隐藏态，获取历史轨迹特征，进而实现自车与周围多车的时间空间关系的交互性建模，并输出车辆未来的预测轨迹。使用美国的NGSIM和德国的High-D自然驾驶车辆轨迹数据集对模型进行训练与测试，并对模型的精度、鲁棒性和迁移性（普适性）进行验证。研究结果表明：与传统模型的预测方法相比，考虑多车交互信息的DIP-LSTM网络的预测方法在预测精度与长时域预测上具有优势，且模型具有良好的迁移性和鲁棒性，显著提高了车辆轨迹预测模型的实用性和普适性。

关键词: 车辆工程, 自动驾驶, 轨迹预测, 多车交互, 长短期记忆神经网络

Abstract:

A DIP-LSTM model with dynamic interactive poling layer is proposed， which enables neighboring vehicles to share hidden states of LSTM network by pooling to get the characteristic of historical trajectory， and then realizes interactive modeling of time-space relationship between target vehicle and surrounding vehicles. NGSIM from USA and High-D from Germany are used to train and test the model， and the accuracy， robustness and transferability of the model are verified. The results show that compared with the traditional model prediction method， the DIP-LSTM network show advantages in prediction accuracy and long-time prediction considering multi-vehicle interactive information， and the model has good transferability and robustness， which significantly improves the practicability and universality of intelligent vehicle trajectory prediction model.

Key words: vehicle engineering, automatic driving, trajectory prediction, multi-vehicle interaction, long short-term memory

中图分类号:

U495

黄玲,崔躜,游峰,洪佩鑫,钟浩川,曾译萱. 适用于多车交互场景的车辆轨迹预测模型[J]. 吉林大学学报(工学版), 2024, 54(5): 1188-1195.

Ling HUANG,Zuan CUI,Feng YOU,Pei-xin HONG,Hao-chuan ZHONG,Yi-xuan ZENG. Vehicle trajectory prediction model for multi-vehicle interaction scenario[J]. Journal of Jilin University(Engineering and Technology Edition), 2024, 54(5): 1188-1195.

图/表 8

图1

图2

图3

表1

图4

图5

表2

表3

参考文献 19

1	Lefevre S, Vasquez D, Laugier C. A survey on motion prediction and risk assessment for intelligent vehicles[J]. Robomech Journal, 2014, 1(1):No.101186.
2	Ammoun S, Nashashibi F. Real time trajectory prediction for collision risk estimation between vehicles[C]∥ IEEE International Conference on Intelligent Computer Communication & Processing, Cluj-Napoca, Romania, 2009: 417-422.
3	Broadhurst A, Baker S, Kanade T. Monte Carlo road safety reasoning[C]∥ 2005 IEEE Intelligent Vehicles Symposium Proceedings, Las Vegas, USA, 2005: 319-324.
4	Aoude G S, Desaraju V R, Stephens L H, et al. Driver behavior classification at intersections and validation on large naturalistic data set[J]. IEEE Transactions on Intelligent Transportation Systems, 2012, 13(2):724-736.
5	Xin L, Wang P, Chan C, et al. Intention-aware long horizon trajectory prediction of surrounding vehicles using dual LSTM networks[C]∥ International IEEE Conference on Intelligent Transportation Systems (ITSC), Maui, USA, 2018: 1441-1446.
6	季学武, 费聪, 何祥坤, 等. 基于LSTM网络的驾驶意图识别及车联轨迹预测[J]. 中国公路学报, 2019, 32(6): 34-42.
	Ji Xue-wu, Fei Cong, He Xiang-kun, et al. Intention recognition and trajectory prediction for vehicles LSTM networks[J]. China Journal of Highway and Transport, 2019, 32(6): 34-42.
7	Deo N, Trivedi M M. Multi-Modal trajectory prediction of surrounding vehicles with maneuver based LSTMs[C]∥ IEEE Conference on Intelligent Vehicles (IV), Changshu, China, 2018: 1179-1184.
8	Altché F, Fortelle A D L. An LSTM network for highway trajectory prediction[C]∥ International IEEE Conference on Intelligent Transportation Systems (ITSC), Yokohama, Japan, 2017: 353-359.
9	Deo N, Trivedi M M. Convolutional social pooling for vehicle tajectory prediction[C]∥ International Conference on Computer Vision and Pattern Recognition (CVPR), Salt Lake, USA, 2018: 1549-1557.
10	Zhang H, Wang Y, Liu J, et al. A multi-modal states based vehicle descriptor and dilated convolutional social pooling for vehicle trajectory prediction[C]∥ SAE Technical Papers, New York, USA, 2020:1-8.
11	Wang J Q, Huang H Y, Li K Q, et al. Towards the unified principles for level 5 autonomous vehicles[J/OL]. [2022-06-10].
12	Sajjad M, Omar Y, Jarrah A, et al. Deep learning-based vehicle behavior prediction for autonomous driving applications: a review[J]. IEEE Transactions on Intelligent Transportation Systems, 2020, 23(1): 33-47.
13	Alahi A, Goel K, Ramanathan V, et al. Social LSTM: human trajectory prediction in crowded spaces[C]∥ IEEE Conference on Computer Vision and Pattern Recognition(CVPR), New York, USA, 2016: 961-971.
14	Graves A. Generating sequences with recurrent neural networks[J]. arXiv Preprint, 2013, 8:No.13080850.
15	Federal Highway Administration. NGSIM-Next generation simulation [EB/OL]. [2022-06-10]. .
16	Robert K, Julian B, Lutz E, et al. The highD dataset: a drone dataset of naturalistic vehicle trajectories on german highways for validation of highly automated driving systems [C]∥ International IEEE Conference on Intelligent Transportation Systems (ITSC), Maui, USA, 2018: 2118-2125.
17	Savitzky A, Golay M J. Smoothing and differentiation of data by simplified least squares procedures[J]. Analytical Chemistry, 1964, 36(8): 1627-1639.
18	黄玲, 郭亨聪, 张荣辉, 等. 人机混驾环境下基于LSTM的无人驾驶车辆换道行为模型[J]. 中国公路学报, 2020, 33(7): 156-165.
	Huang Ling, Guo Heng-cong, Zhang Rong-hui, et al. LSTM-based lane-changing behavior model for unmanned vehicle under environment of heterogeneous human-driven and autonomous vehicles[J]. China Journal of Highway and Transport, 2020, 33(7): 156-165.
19	Huber P J. Robust statistics[M]. New York: John Wiley & Son, 2004.

相关文章 15

[1]	郭洪艳,王连冰,赵旭,戴启坤. 考虑侧向运动的整车质量与道路坡度估计[J]. 吉林大学学报(工学版), 2024, 54(5): 1175-1187.
[2]	汪少华,张启睿,施德华,殷春芳,李春. 双行星排式混合动力传动系统非线性振动响应特性分析[J]. 吉林大学学报(工学版), 2024, 54(4): 890-901.
[3]	曲大义,张可琨,顾原,王韬,宋慧,戴守晨. 自动驾驶车辆换道决策行为分析及分子动力学建模[J]. 吉林大学学报(工学版), 2024, 54(3): 700-710.
[4]	高镇海,蔡荣贵,孙天骏,于桐,赵浩源,班浩. 人机共驾下的驾驶行为数据滤波方法[J]. 吉林大学学报(工学版), 2024, 54(3): 589-599.
[5]	谢宪毅,王禹涵,金立生,赵鑫,郭柏苍,廖亚萍,周彬,李克强. 基于改变控制时域时间步长的智能车轨迹跟踪控制[J]. 吉林大学学报(工学版), 2024, 54(3): 620-630.
[6]	涂辉招,王万锦,乔鹏,郭静秋,鹿畅,吴海飞. 自动驾驶卡车路测安全员接管干预行为解析[J]. 吉林大学学报(工学版), 2024, 54(3): 727-740.
[7]	胡宏宇,张慧珺,姚荣涵,陈国迎,高菲. L3级自动驾驶接管过程驾驶员情景意识研究[J]. 吉林大学学报(工学版), 2024, 54(2): 410-418.
[8]	邓小林,杨馥模,覃善甘. 新型仿竹六边形梯度层级多胞管耐撞性对比分析[J]. 吉林大学学报(工学版), 2024, 54(2): 333-345.
[9]	王毅刚,王玉鹏,张昊,赵思安. 高速列车转向架区域气动噪声源识别与分析[J]. 吉林大学学报(工学版), 2024, 54(2): 346-355.
[10]	聂建军,侯军凯,解晓琳,鄢鸿桢. 新型巡检机器人移动底盘设计及越障性能分析[J]. 吉林大学学报(工学版), 2024, 54(2): 356-364.
[11]	王铁,李旭东,田程,赵宏伟. 基于多轴载荷投影构建轮辋双轴疲劳损伤模型[J]. 吉林大学学报(工学版), 2024, 54(1): 99-104.
[12]	李旭东,王新宇,田程,张新峰,牛治慧,赵志强. 基于用户关联的车辆耐久性载荷谱编制[J]. 吉林大学学报(工学版), 2024, 54(1): 66-75.
[13]	吴骁,史文库,郭年程,赵燕燕,陈志勇,李鑫鹏,孙卓,刘健. 基于Ease off的准双曲面齿轮多目标优化[J]. 吉林大学学报(工学版), 2024, 54(1): 76-85.
[14]	陈兆玮,蒲前华. 弹性车轮对大跨斜拉桥车桥耦合振动的抑制特性[J]. 吉林大学学报(工学版), 2023, 53(9): 2519-2532.
[15]	张健,李青扬,李丹,姜夏,雷艳红,季亚平. 基于深度强化学习的自动驾驶车辆专用道汇入引导[J]. 吉林大学学报(工学版), 2023, 53(9): 2508-2518.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

模型	预测时域
模型	1 s	2 s	3 s	4 s	5 s
LSTM	0.68	1.76	3.21	4.82	6.87
S-LSTM	0.63	1.43	2.63	4.12	6.18
SV-LSTM	0.62	1.41	2.44	3.87	5.73
DIP-LSTM	0.58	1.24	2.12	3.14	4.33

干扰	模型	预测时域/s
干扰	模型	1 s	2 s	3 s	4 s	5 s
±10%噪声	LSTM	1.57	4.08	7.25	11.18	16.91
	S-LSTM	1.32	3.15	5.87	9.24	14.23
	SV-LSTM	1.29	3.11	5.81	9.13	14.07
	DIP-LSTM	1.15	2.86	5.43	8.18	11.67
±20%噪声	LSTM	2.23	6.17	11.05	19.27	28.24
	S-LSTM	2.01	5.46	9.56	16.24	23.34
	SV-LSTM	1.97	5.41	9.47	15.82	22.67
	DIP-LSTM	1.85	4.97	8.61	14.17	19.87
±30%噪声	LSTM	3.35	9.25	16.48	31.28	41.25
	S-LSTM	3.08	8.32	14.64	26.32	37.62
	SV-LSTM	3.02	8.25	14.21	25.87	36.58
	DIP-LSTM	2.84	7.73	13.15	23.21	31.18

数据集	模型	RMSE预测时域/s
数据集	模型	1 s	2 s	3 s	4 s	5 s
NGSIM I-101	LSTM	0.86	2.31	4.94	8.36	11.31
	S-LSTM	0.73	1.86	3.76	6.63	8.21
	SV-LSTM	0.71	1.78	3.52	6.29	7.78
	DIP-LSTM	0.64	1.65	2.87	5.24	6.62
High-D	LSTM	2.16	5.57	8.86	15.02	21.12
	S-LSTM	1.83	4.67	7.24	11.56	16.72
	SV-LSTM	1.78	4.55	7.02	11.02	16.14
	DIP-LSTM	1.42	3.25	6.42	9.62	14.27

适用于多车交互场景的车辆轨迹预测模型

Vehicle trajectory prediction model for multi-vehicle interaction scenario

RICH HTML

PDF (PC)

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 19

相关文章 15

Metrics

本文评价

推荐阅读 0