考虑车辆运动状态信息特性的自由换道意图识别模型

doi:10.16097/j.cnki.1009-6744.2025.06.007

交通运输系统工程与信息 ›› 2025, Vol. 25 ›› Issue (6): 74-86.DOI: 10.16097/j.cnki.1009-6744.2025.06.007

• 智能交通系统与信息技术 • 上一篇下一篇

考虑车辆运动状态信息特性的自由换道意图识别模型

辛琪^*，王彦锋，王智龙，王畅，牛世峰

长安大学，汽车学院，西安710021

收稿日期:2025-07-24 修回日期:2025-09-23 接受日期:2025-09-29 出版日期:2025-12-25 发布日期:2025-12-24
作者简介:辛琪(1987—)，男，陕西咸阳人，副教授。
基金资助:
陕西省重点研发计划项目(2024CY2-GJHX-87)；长安大学中央高校基本科研业务费专项资金(300102223205)。

A Free Lane Change Intention Recognition Model Considering Vehicle Motion State Information Characteristics

XIN Qi^*, WANG Yanfeng, WANG Zhilong, WANG Chang, NIU Shifeng

School of Automobile, Chang'an University, Xi'an 710021, China

Received:2025-07-24 Revised:2025-09-23 Accepted:2025-09-29 Online:2025-12-25 Published:2025-12-24
Supported by:
Key Research and Development Program Project of Shaanxi Province(2024CY2-GJHX-87)； Fundamental Research Funds for the Central Universities, CHD(300102223205)。

摘要/Abstract

摘要： 为实现自由换道场景下驾驶人换道意图的准确识别，通过分析换道过程中车辆运动状态的信息变化规律，提出一种考虑车辆运动状态信息特性的自由换道意图识别模型。首先，基于人机共驾实车系统平台采集驾驶人自由换道场景下的车辆运动状态信息，通过车道线检测获取车辆中心点与车道中心线的偏离距离，确定自由换道过程关键时间节点，将所采集数据分成车道保持、向左换道和向右换道3类，构建换道意图数据集；其次，通过SHAP(SHapley Additive exPlanations)全局可解释性方法分析各车辆运动状态信息对换道意图识别的影响权重，并采用独立样本T检验说明各变量的差异性，验证各变量作为换道意图识别模型输入的可行性；再次，针对自由换道过程中车辆运动状态的非平稳性、采样吉布斯现象及车道偏离数据干扰问题，在Informer网络的基础上依次引入可逆实例归一化模块(RevIN)、频率增强信道注意力机制(FECAM)、趋势感知自注意力机制(ETTA)和Unet结构，构建自由换道意图识别模型RF-EUInformer(RevINFECAM-ETTA Unet Informer)；最后，采用蒙特卡洛交叉验证评估模型的泛化能力，通过消融试验得出引入的各模块在1.5s预判时间下对准确率的贡献分别为1.4%、0.5%、0.6%和1.2%，验证了各模块的有效性。与双向长短期记忆网络(Bi-LSTM)、卷积长短期记忆网络(ConvLSTM)、时域卷积网络(TCN)和时域卷积网络注意力机制(TCN-Attention)等模型进行对比分析，所提出模型在0.5、1.0、1.5s预判时间下的准确率较最优对比模型分别提升了3.9%、4.5%和5.8%。

关键词: 智能交通, 自由换道意图识别, Informer, 车辆运动状态信息, SHAP方法

Abstract: To accurately identify the driver's lane changing intention in the scene of free lane changing, this paper proposes a free lane changing intention recognition model considering the characteristics of vehicle's motion state information by analyzing the change law of vehicle's motion state information in lane changing. First, the vehicle movement status information under free lane changing was collected based on the human-machine co-driving real vehicle system platform. The distance between the vehicle center and the lane center was obtained through lane detection to determine the key time nodes of the free lane changing process. The collected data were divided into three categories: lane keeping, left lane changing and right lane changing, and the lane changing intention dataset was constructed. Then, the influence weight of vehicle motion state information on lane change intention recognition is analyzed using the SHAP (SHapley Additive exPlanations) global interpretability method, and the difference of each variable is illustrated by the independent sample T test, which verifies the feasibility of each variable as the input of lane change intention recognition model. To address challenges in recognizing free lane-changing intentions: such as unstable vehicle movement, the Gibbs phenomenon in data sampling, and interference from lane departure data, a new model was built based on the Informer network, incorporating several key techniques: RevIN (Reversible Instance Normalization), FECAM (Frequency Enhanced Channel Attention Mechanism), ETTA (Efficient Temporal Trend-Aware Attention mechanism), and U-Net structure. This model is named RF-EUInformer (RevIN FECAM-ETTA Unet Informer). The Monte Carlo Cross Validation was used to evaluate the generalization ability of the model. The ablation test showed that the contribution of each module to the accuracy was respectively 1.4%, 0.5%, 0.6% and 1.2% in 1.5 s prediction time, which verified the effectiveness of each module. Compared with Bi-LSTM (Bidirectional Long Short Term Memory), ConvLSTM (Convolutional Long Short-Term Memory), TCN (Temporal Convolutional Network) and TCN-Attention models, the accuracy of the proposed model in 0.5 s, 1.0 s and 1.5 s prediction time is improved by 3.9%, 4.5% and 5.8%, respectively.

Key words: intelligent transportation, free lane change intention recognition, Informer, vehicle movement status information, SHAP method

中图分类号:

U491.2

辛琪, 王彦锋, 王智龙, 王畅, 牛世峰. 考虑车辆运动状态信息特性的自由换道意图识别模型[J]. 交通运输系统工程与信息, 2025, 25(6): 74-86.

XIN Qi, WANG Yanfeng, WANG Zhilong, WANG Chang, NIU Shifeng. A Free Lane Change Intention Recognition Model Considering Vehicle Motion State Information Characteristics[J]. Journal of Transportation Systems Engineering and Information Technology, 2025, 25(6): 74-86.

导出引用管理器 EndNote|Ris|BibTeX

链接本文: http://www.tseit.org.cn/CN/10.16097/j.cnki.1009-6744.2025.06.007

http://www.tseit.org.cn/CN/Y2025/V25/I6/74

参考文献

[1]ZHANG Y, XU Q, WANG J, et al. A learning-based discretionary lane-change decision-making model with driving style awareness[J]. IEEE Transactions on Intelligent Transportation Systems, 2022, 24(1): 68-78.

[2]付锐,张海伦,刘文晓,等.驾驶人意图识别综述[J].长安大学学报(自然科学版),2022,42(1): 33-60. [FU R, ZHANG H L, LIU W X, et al. Review on driver intention recognition[J]. Journal of Chang'an University(Natural Science Edition), 2022, 42(1): 33-60.]

[3]王婉琦,程国柱,徐亮.基于无监督聚类分析的激进换道行为识别方法[J].交通运输系统工程与信息,2024, 24(2): 166-178. [WANG W Q, CHENG G Z, XU L. Identification of aggressive lane-changing behaviour based on unsupervised cluster analysis[J]. Journal of Transportation Systems Engineering and Information Technology, 2024, 24(2): 166-178.]

[4]卢辉遒,赵枫,谢波,等.冰雪环境下基于神经网络的驾驶人换道意图识别[J]. 吉林大学学报(工学版), 2023, 53(1): 273-284. [LU H Q, ZHAO F, XIE B, et al. Driver's lane change intention recognition in snow and ice environment based on neural network[J]. Journal of Jilin University(Engineering and Technology Edition), 2023, 53(1): 273-284.]

[5] FENG Z, WEI X, BI Y, et al. An integrated framework for driving risk evaluation that combines lane-changing detection and an attention-based prediction model[J]. Traffic Injury Prevention, 2025, 26(2): 198-206.

[6]赵建东,赵志敏,屈云超,等.轨迹数据驱动的车辆换道意图识别研究[J].交通运输系统工程与信息,2022, 22(4): 63-71. [ZHAO J D, ZHAO Z M, QU Y C, et al. Vehicle lane change intention recognition driven by trajectory data[J]. Journal of Transportation Systems Engineering and Information Technology, 2022, 22(4): 63-71.]

[7] 李琳,赵万忠,王春燕.基于Bi-GLSTM网络的车辆驾驶意图分析与识别[J]. 机械工程学报,2024, 60(10): 51-63. [LI L, ZHAO W Z, WANG C Y. Driving intention recognition model based on Bi-GLSTM network[J]. Journal of Mechanical Engineering, 2024, 60(10): 51 63.]

[8]LIU H, WANG T, LI W, et al. Lane-change intention recognition considering oncoming traffic: Novel insights revealed by advances in deep learning[J]. Accident Analysis & Prevention, 2024, 198: 107476.

[9] GAO K, LI X, CHEN B, et al. Dual transformer based prediction for lane change intentions and trajectories in mixed traffic environment[J]. IEEE Transactions on Intelligent Transportation Systems, 2023, 24(6): 6203 6216.

[10] 苑仁腾, 王晨竹,项乔君,等.轨迹数据驱动的车辆换道意图识别模型[J].华南理工大学学报(自然科学版), 2024, 52(6): 34-44. [YUAN R T, WANG C Z, XIANG Q J, et al. Trajectory data-driven model for vehicle lane change intention recognition[J]. Journal of South China University of Technology(Natural Science Edition), 2024, 52(6): 34-44.]

[11] 杨达, 刘家威,郑斌,等.基于时域卷积网络与注意力机制的车辆换道轨迹预测模型[J].交通运输系统工程与信息,2024, 24(2): 114-126. [YANG D, LIU J W, ZHENG B, et al. A vehicle lane-changing trajectory prediction model based on temporal convolutional networks and attention mechanism[J]. Journal of Transportation Systems Engineering and Information Technology, 2024, 24(2): 114-126.]

[12] LUNDBERG S M, LEE S I. A unified approach to interpreting model predictions[J]. Advances in Neural Information Processing Systems, 2017, 30: 1-10.

[13] ZHOU H, ZHANG S, PENG J, et al. Informer: Beyond efficient transformer for long sequence time-series forecasting[C]// 35th AAAI Conference on Artificial Intelligence/33rd Conference on Innovative Applications of Artificial Intelligence/11th Symposium on Educational Advances in Artificial Intelligence, Palo Alto: AAAI, 2021: 11106-11115.

[14] LI Y, LIU F, XING L, et al. A deep learning framework to explore influences of data noises on lane-changing intention prediction[J]. IEEE Transactions on Intelligent Transportation Systems, 2024, 25(7): 6514-6526.

[1]	杨海飞, 唐勇, 郭延永, 李红伟, 赵恩泽. 可接受间距策略下网联车辆队列安全优化控制方法[J]. 交通运输系统工程与信息, 2025, 25(6): 50-61.
[2]	何永明, 卢杨彭, 刘汇洋, 李鑫然. 超高速公路智能网联车辆编队策略研究[J]. 交通运输系统工程与信息, 2025, 25(6): 62-73.
[3]	李雪玮, 孙齐, 刘筱萌, 康学建, 赵晓华. 基于冲突理论的平视显示行人预警系统安全效用建模分析[J]. 交通运输系统工程与信息, 2025, 25(6): 350-359.
[4]	田大新, 肖啸, 周建山. AI驱动的自动驾驶汽车轨迹预测方法综述[J]. 交通运输系统工程与信息, 2025, 25(5): 1-24.
[5]	程国柱, 王文志, 杨子涵, 汪国鹏, 陈永胜, 顾爽. 网联自动驾驶车辆专用车道设置研究综述[J]. 交通运输系统工程与信息, 2025, 25(5): 25-39.
[6]	吴付威, 杨慧, 马勇, 崔同川, 闫晟煜, 张智. L3级自动驾驶车辆两阶段预警对接管绩效影响研究[J]. 交通运输系统工程与信息, 2025, 25(5): 40-49.
[7]	窦雪萍, 杨明慧, 熊杰. 考虑自适应控制点的自动驾驶公交运行优化方法[J]. 交通运输系统工程与信息, 2025, 25(5): 50-58.
[8]	宋浪, 胡晓伟, 俞山川, 安实. 网联自动驾驶环境下交叉口车道分配与车辆轨迹协同优化[J]. 交通运输系统工程与信息, 2025, 25(5): 59-71.
[9]	王殿海, 王奕飞, 黄宇浪, 刘泳, 曾佳棋. 车牌识别设备稀疏布局下的路段行程时间估计方法[J]. 交通运输系统工程与信息, 2025, 25(5): 83-90.
[10]	林培群, 黄超铄, 周楚昊, 庞崇浩, 邓锴宇. 数据驱动的高速公路自学习元胞传输模型[J]. 交通运输系统工程与信息, 2025, 25(5): 103-113.
[11]	李传耀, 章孟丹, 唐铁桥. 考虑模块化公交供应限制的双模式早通勤均衡分析[J]. 交通运输系统工程与信息, 2025, 25(5): 205-214.
[12]	沈鹏举, 宋丽英, 林川倩. 高铁列车晚点数据的异常检测与增量聚类方法[J]. 交通运输系统工程与信息, 2025, 25(5): 236-247.
[13]	杨静, 侯宇晴, 谢余晨, 杨安安. 城市轨道交通网络短时OD客流多任务协同预测[J]. 交通运输系统工程与信息, 2025, 25(5): 261-270.
[14]	姚振兴, 刘贤, 赵一飞, 王亮, 王彦琛. 手机信令不均匀定位下出行端点自适应识别方法[J]. 交通运输系统工程与信息, 2025, 25(4): 44-52.
[15]	张鹏, 李兴旺, 姬炳豪, 孙超, 李文权. 路口重复放行的公交与社会车辆协同绿波优化模型[J]. 交通运输系统工程与信息, 2025, 25(4): 53-62.

考虑车辆运动状态信息特性的自由换道意图识别模型

A Free Lane Change Intention Recognition Model Considering Vehicle Motion State Information Characteristics

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics