广西师范大学学报(自然科学版) ›› 2026, Vol. 44 ›› Issue (6): 13-26.doi: 10.16088/j.issn.1001-6600.2026033103

• 智慧交通 •    下一篇

基于DBSCAN与注意力BiLSTM的驾驶行为偏好识别研究

姜宏飞, 邴其春*, 卢元锋, 邢丞思, 马宇航   

  1. 青岛理工大学 智慧交通与运载学院, 山东 青岛 266520
  • 收稿日期:2026-03-31 修回日期:2026-06-02 出版日期:2026-11-05 发布日期:2026-11-10
  • 通讯作者: 邴其春(1989—),男,山东即墨人,青岛理工大学副教授,博士。E-mail: bingqichun@163.com
  • 基金资助:
    山东省自然科学基金(ZR2023MG058)

Research on driving behavior preference recognition based on DBSCAN and attention BiLSTM

Jiang Hongfei, Bing Qichun*, Lu Yuanfeng, Xing Chengsi, Ma Yuhang   

  1. Qingdao University of Technology, School of Intelligent Transportation and Vehicle Engineering, Qingdao Shandong 266520, China
  • Received:2026-03-31 Revised:2026-06-02 Online:2026-11-05 Published:2026-11-10

摘要: 针对高速公路场景下驾驶行为偏好难以准确识别的问题,本文基于NGSIM US-101车辆轨迹数据,提取速度、加速度、加加速度、车头时距、车头间距及换道等特征,采用基于密度的空间聚类算法(DBSCAN)结合驾驶激进度评分构建谨慎型、一般型和激进型3类驾驶行为偏好标签。在此基础上,建立融合注意力机制的双向长短期记忆网络(BiLSTM)识别模型,对驾驶行为序列进行分类。结果表明,2 490辆有效车辆中,3类样本占比分别约为26.1%、44.7%和29.1%;不同类别在速度水平、纵向操控强度及跟驰安全裕度等特征上存在显著差异;所建模型对驾驶行为偏好的识别准确率达到90.25%。该方法可为智能网联车辆的行为预测、决策规划与安全控制提供参考。

关键词: 交通运输工程, 驾驶行为偏好, DBSCAN算法, 注意力机制, BiLSTM模型, 监督学习

Abstract: To address the difficulty of accurately identifying driving behavior preferences in highway scenarios, vehicle trajectory data from NGSIM US-101are used in this paper.The data features including speed, acceleration, jerk, time headway, spacing, and lane-changing behavior are extracted. A three-class labeling method for driving behavior preferences, namely cautious, normal, and aggressive, is constructed by combining the density-based spatial clustering of applications with noise (DBSCAN) algorithm with an aggressiveness score. On this basis, a bidirectional long short-term memory (BiLSTM) identification model integrated with an attention mechanism is established to classify driving behavior sequences. The results show that, among 2 490 valid vehicles, the proportions of the three classes are 26.1%, 44.7%, and29.1%, respectively. Significant differences are observed among the classes in terms of speed level, longitudinal control intensity, and car-following safety margin. The proposed model achieves an accuracy of 90.25% in driving behavior preference identification. This method provides a reference for behavior prediction, decision-making and planning, and safety control of intelligent connected vehicles.

Key words: transportation engineering, driving behavior preference, DBSCAN algorithm, attention mechanism, BiLSTM model, supervised learning

中图分类号:  U495

[1] Zhang Y T, Wu S Q, Mubbala D, et al. Human-machine cooperation strategies in game-theoretic scenarios within mixed traffic: a simulator study on driving styles[J]. Accident Analysis & Prevention, 2025, 223: 108254. DOI:10.1016/j.aap.2025.108254.
[2] Li S B, Hu X H, et al. Lane-changing prediction model of autonomous vehicle based on Bayesian game[J]. Journal of Measurement Science and Instrumentation, 2023, 14(2): 242-252. DOI:10.62756/jmsi.1674-8042.2023028.
[3] Cao Y X, Jiang Y N, Zeng X R. Adaptive game-theoretic decision-making with driving style recognition for autonomous vehicles in uninterrupted traffic flows at intersections[J]. Robotics and Autonomous Systems, 2025, 194: 105180. DOI:10.1016/j.robot.2025.105180.
[4] Elander J, West R, French D. Behavioral correlates of individual differences in road-traffic crash risk: an examination method and findings[J]. Psychological Bulletin, 1993, 113(2): 279-294. DOI:10.1037/0033-2909.113.2.279.
[5] Li X, Yang B, Zeng X M, et al. StyleFormer: multi-agent joint trajectory prediction and planning in urban environments with driving style awareness[J]. IEEE Transactions on Intelligent Transportation Systems, 2025, 26(11): 20317-20333. DOI:10.1109/TITS.2025.3595733.
[6] He Y, Hu Y R, Li J P, et al. Multi-scenario driving style research based on driving behavior pattern extraction[J]. Accident Analysis & Prevention, 2025, 214: 107963. DOI:10.1016/j.aap.2025.107963.
[7] 刘思源, 喻伟, 刘洁莹, 等. 考虑驾驶风格的车辆换道行为及预测模型[J]. 长沙理工大学学报(自然科学版), 2019, 16(1): 28-35. DOI:10.3969/j.issn.1672-9331.2019.01.005.
[8] 李立治, 杨建军, 刘双喜, 等. 国内人群的驾驶风格分类及识别方法研究[J]. 重庆理工大学学报(自然科学), 2019, 33(11): 33-40.
[9] 万豫, 黄妙华, 王思楚. 基于改进DBSCAN算法的驾驶风格识别方法研究[J]. 合肥工业大学学报(自然科学版), 2020, 43(10): 1313-1320. DOI:10.3969/j.issn.1003-5060.2020.10.003.
[10]Li B X, Liu Y F, Zhao J W, et al. Representing-behavior-based online driving style recognition and its road verification[J]. Engineering Applications of Artificial Intelligence, 2025, 156: 111241. DOI:10.1016/j.engappai.2025.111241.
[11]Xu T, Wu K, Zhu Y D, et al. Driving style recognition at first impression for online trajectory prediction[J]. IFAC-PapersOnLine, 2023, 56(2): 11287-11292. DOI:10.1016/j.ifacol.2023.10.323.
[12]Hu X H, Chen S Z, Zhao J H, et al. Risk identification and prediction model for continuous-lane-change vehicles considering driving style[J]. Expert Systems with Applications, 2025, 259: 125292. DOI:10.1016/j.eswa.2024.125292.
[13]王兴鸿, 鲁燕, 高雄. 基于MTL-MTF-LSTM神经网络的驾驶风格识别方法[J]. 汽车实用技术, 2025, 50(3): 49-53. DOI:10.16638/j.cnki.1671-7988.2025.003.009.
[14]梁科, 陈华晟, 潘明章, 等. 采用双向LSTM自编码器的驾驶风格谱聚类识别研究[J]. 重庆理工大学学报(自然科学), 2023, 37(10): 28-37.
[15]靳引利, 刘语欣, 王华, 等. 基于轨迹数据的高速公路换道行为分类与辨识方法[J]. 长安大学学报(自然科学版), 2026, 46(3): 1-12. DOI: 10.19721/j.cnki.1671-8879.2026.03.001.
[16]李俊涛, 张明恒, 王潇雨, 等. 考虑驾驶风格的驾驶意图识别模型[J]. 大连理工大学学报, 2026, 66(2): 147-155. DOI:10.7511/dllgxb202602006.
[17]周卫琪, 杨磊, 施德华, 等. 基于驾驶风格的横纵向耦合风险场建模及灵敏度分析[J/OL]. 汽车工程学报: 1-13[2026-05-22]. https://kns.cnki.net/kcms/detail/detail.aspx?dbcode=CJFD&dbname=CJFD&filename=QCYK20260511002.
[18]Mei P, Karimi H R, Ou L, et al. Driving style classification and recognition methods for connected vehicle control in intelligent transportation systems: a review[J]. ISA Transactions, 2025, 158: 167-183. DOI:10.1016/j.isatra.2025.01.033.
[19]Messaoud K, Yahiaoui I, Verroust-Blondet A, et al. Attention based vehicle trajectory prediction[J]. IEEE Transactions on Intelligent Vehicles, 2021, 6(1): 175-185. DOI:10.1109/TIV.2020.2991952.
[20]杨玉凤, 曲大义, 兰添贺, 等. 基于机器学习的个性化驾驶风格倾性识别方法[J/OL]. 深圳大学学报(理工版): 1-9[2024-01-04]. https://kns.cnki.net/kcms/detail/detail.aspx?dbcode=CJFD&dbname=CJFD&filename=SZDL20240102001.
[1] 黄泽华, 李泽龙, 周绍园, 邱芹军, 龚秉生, 梁栋, 马凯. 联合物理信息与深度学习的滑坡易发性评价方法[J]. 广西师范大学学报(自然科学版), 2026, 44(6): 54-68.
[2] 钱俊磊, 王熹之, 曾凯, 杜学强, 刘贺, 朱立光. 基于MHTD-YOLO11n的钢材表面缺陷检测算法[J]. 广西师范大学学报(自然科学版), 2026, 44(3): 60-74.
[3] 黄艳国, 肖洁, 吴水清. 基于D2STGNN的双向高效多尺度交通流预测[J]. 广西师范大学学报(自然科学版), 2026, 44(1): 10-22.
[4] 刘志豪, 李自立, 苏珉. 智能通信与无人机结合的YOLOv8电动车骑行者头盔佩戴检测方法[J]. 广西师范大学学报(自然科学版), 2026, 44(1): 23-32.
[5] 施子豪, 蒙祖强, 谈超洪. 基于注意力机制和多尺度融合的多模态虚假新闻检测模型[J]. 广西师范大学学报(自然科学版), 2026, 44(1): 68-79.
[6] 黎豊玮, 谭玉枚, 宋树祥, 夏海英. 基于注意力引导的遮挡感知面部表情识别[J]. 广西师范大学学报(自然科学版), 2025, 43(5): 104-113.
[7] 田晟, 熊辰崟, 龙安洋. 基于改进PointNet++的城市道路点云分类方法[J]. 广西师范大学学报(自然科学版), 2025, 43(4): 1-14.
[8] 韩烁, 江林峰, 杨建斌. 基于注意力机制PINNs方法求解圣维南方程[J]. 广西师范大学学报(自然科学版), 2025, 43(4): 58-68.
[9] 石天怡, 南新元, 郭翔羽, 赵濮, 蔡鑫. 基于改进ConvNeXt的苹果叶片病害分类算法[J]. 广西师范大学学报(自然科学版), 2025, 43(4): 83-96.
[10] 卢展跃, 陈艳平, 杨卫哲, 黄瑞章, 秦永彬. 基于掩码注意力与多特征卷积网络的关系抽取方法[J]. 广西师范大学学报(自然科学版), 2025, 43(3): 12-22.
[11] 郭翔羽, 石天怡, 陈燕楠, 南新元, 蔡鑫. 基于YOLO-CDBW模型的列车接触网异物检测研究[J]. 广西师范大学学报(自然科学版), 2025, 43(2): 56-69.
[12] 苏春海, 夏海英. 抗噪声双约束网络的面部表情识别[J]. 广西师范大学学报(自然科学版), 2025, 43(2): 70-82.
[13] 刘玉娜, 马双宝. 基于改进YOLOv8n的轻量化织物疵点检测算法[J]. 广西师范大学学报(自然科学版), 2025, 43(2): 83-94.
[14] 戴林华, 黎远松, 石睿, 何忠良, 李雷. HSED-YOLO:一种轻量化的带钢表面缺陷检测模型[J]. 广西师范大学学报(自然科学版), 2025, 43(2): 95-106.
[15] 余快, 宋宝贵, 邵攀, 余翱. 基于层级尺度交互的U-Net遥感影像建筑物提取方法[J]. 广西师范大学学报(自然科学版), 2025, 43(2): 121-132.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!
版权所有 © 广西师范大学学报(自然科学版)编辑部
地址:广西桂林市三里店育才路15号 邮编:541004
电话:0773-5857325 E-mail: gxsdzkb@mailbox.gxnu.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发