Journal of Jilin University(Engineering and Technology Edition) ›› 2026, Vol. 56 ›› Issue (8): 2084-2094.doi: 10.13229/j.cnki.jdxbgxb.20250008

Previous Articles    

Vehicle lane change intention recognition based on PSO⁃GRU model

Rui-jun GUO(),Chao-ran FAN,Ming-di FU   

  1. School of Transportation Engineering，Dalian Jiaotong University，Dalian 116028，China
  • Received:2025-01-04 Online:2026-08-01 Published:2026-09-02

RICH HTML

 0   

PDF (PC)

4

Abstract:

To address the issues of low efficiency and limited solution space caused by manual hyperparameter tuning when using traditional neural network models for vehicle lane-changing intention recognition， a lane-changing intention recognition model based on PSO-GRU was proposed. First， a multi-step trajectory smoothing method was developed considering the characteristics of horizontal and vertical trajectories in the NGSIM dataset. Next， the trajectory data were mined for parameters and reconstructed into the dataset to establish a vehicle lane-changing trajectory database. The importance of features in the database was evaluated using the random forest algorithm. Finally， a combined model for lane-changing intention recognition based on PSO-GRU was built， with the PSO algorithm used to optimize the hyperparameters of the GRU network to find the optimal configuration. The results demonstrate that， compared to traditional deep learning and machine learning algorithms， the proposed PSO-GRU model exhibits higher recognition accuracy and more stable performance.

Key words: engineering of communications and transportation system, lane change intention recognition, multi-step trajectory smoothing, random forest algorithm, deep learning, particle swarm optimization algorithm

CLC Number: 

  • U495

Table 1

Main parameter of NGSIM"

字段说明单位及描述
Vehicle_ID车辆编号-
Frame_ID数据帧号1/10 s
Local_X采集区域车辆的横坐标ft
Local_Y采集区域车辆的纵坐标ft
v_Class车辆类型1-摩托车；2-小汽车；3-大型车
v_Vel车辆速度ft/s
v_Acc车辆加速度ft/s2
Lane_ID所在车道-
Preceding跟驰前车编号-
Following跟驰后车编号-
Location路段名称-

Fig.1

Comparison of lateral data smoothing effects"

Fig.2

Comparison of smoothing windows"

Fig.3

Comparison of longitudinal data smoothing effects"

Fig.4

Statistics of vehicle lateral offset"

Fig.5

Schematic diagram of associated vehicles"

Table 2

Input parameter indicators"

特征名称特征描述

自身

车辆

Local_X车辆横坐标/m
Local_Y车辆纵坐标/m
vx车辆横向速度/（m·s-1）
vy车辆纵向速度/（m·s-1）
ax车辆横向加速度/（m·s-2）
ay车辆纵向加速度/（m·s-2）
angle车辆偏角/（°）

关联

车辆

FY前车纵向位置差/m
FV前车纵向速度差/（m·s-1）
BY后车纵向位置差/m
BV后车纵向速度差/（m·s-1）
LFY左前车纵向位置差/m
LFV左前车纵向速度差/（m·s-1）
LBY左后车纵向位置差/m
LBV左后车纵向速度差/（m·s-1）
RFY右前车纵向位置差/m
RFV右前车纵向速度差/（m·s-1）
RBY右后车纵向位置差/m
RBV右后车纵向速度差/（m·s-1）

Table 3

Random forest classification accuracy"

树的数量未标准化分类准确率/%标准化分类准确率/%增幅/%
10053.2460.106.86
11053.0860.247.16
12052.9360.537.60
13052.6560.227.57
14053.0860.337.25
15053.0260.397.37

Fig.6

Feature importance assessment based on random forest algorithm"

Fig.7

GRU lane change intention recognition model"

Fig.8

PSO-GRU lane change intention recognition model framework"

Fig.9

Comparison of training iteration charts for unsmooth and smooth data"

Table 4

Model recognition metrics"

类别准确率/%精确率/%召回率/%
未平滑90.5991.6590.38
平滑后92.3592.9492.47

Table 5

Optimization of GRU lane change intention recognition model network structure"

参数实验设定数值
输入层InputSize19
GRU层数GruLayer2
隐藏神经元数量NumHiddenUnits50
批次大小batch size128
学习率learning rate0.001
最大迭代次数MaxEpochs100
输出层OutputSize3

Fig.10

GRU lane change intention recognition model confusion matrix"

Table 6

GRU lane change intention recognition model evaluation metrics"

换道意图精确率/%召回率/%F1/%准确率/%
向左换道93.88100.0096.8492.35
保持直行95.7580.3687.38
向右换道89.1997.0692.96

Table 7

PSO-GRU vehicle lane change intention recognition model parameters"

参数数值及描述
种群规模5
搜索维度3
认知因子2
社交因子2
惯性权重［0.4，0.9］
终止条件迭代20次

Fig.11

Iteration chart of particle loss values forPSO-GRU model"

Table 8

PSO-GRU optimal network architecture"

参数实验设定数值
输入层InputSize19
GRU层数GruLayer2
隐藏神经元数量NumHiddenUnits76
批次大小batch size64
学习率learning rate0.01
最大迭代次数MaxEpochs100
输出层OutputSize3

Fig.12

PSO-GRUGRU lane-change intention recognition model confusion matrix"

Table 9

GRU lane-change intention recognition model performance evaluation metrics"

换道意图精确率/%召回率/%F1/%准确率/%
向左换道97.8397.8397.8397.06
保持直行94.7496.4395.58
向右换道98.5197.0697.78

Fig.13

Model accuracy boxplot"

Table 10

Comparison of performance indicators for various models"

算法模型准确率/%
传统机器学习RF算法60.53
KNN69.09
BP84.36
决策树86.33
深度学习BiLSTM91.76
LSTM92.35
GRU92.35
PSO-GRU97.06
[1] 张海伦, 付锐. 高速场景相邻前车驾驶行为识别及意图预测[J].交通运输系统工程与信息, 2020, 20(1): 40-46.
Zhang Hai-lun, Fu Rui.Driving behavior recognition and intention prediction of adiacent preceding vehicle in highway scene[J].Journal of Transportation Systems Engineering and Information Technology, 2020, 20(1): 40-46.
[2] David R, Rothe S, Soffker D. Lane changing behavior recognition based on artificial neural network-based state machine approach[C]∥Proceedings of lEEE Conference on Intelligent Transportation Systems, Indianapolis, Indiana, USA, 2021: 3444-3449.
[3] 季学武, 费聪, 何祥坤, 等.基于LSTM网络的驾驶意图识别及车辆轨迹预测[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 using LSTM network[J].China Journal of Highway and Transport, 2019, 32(6): 34-42.
[4] 宋晓琳, 曾艳兵, 曹昊天, 等.基于长短期记忆网络的换道意图识别方法[J].中国公路学报, 2021, 34(11): 236-245.
Song Xiao-lin, Zeng Yan-bing, Cao Hao-tian,et al.Lane change intention recognition method based on an LSTM network[J].China Journal of Highway and Transport, 2021, 34(11): 236-245.
[5] 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 Jian-dong, Zhao Zhi-min, Qu Yun-chao, 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.
[6] 张洪加, 郭应时, 高松, 等. 网联与非网联环境下驾驶人换道意图识别研究[J].中国公路学报,2023,36(9): 257-270.
Zhang Hong-jia, Guo Ying-shi, Gao Song,et al.Driver lane-changing intention recognition in connected and non-connected environments[J].China Journal of Highway and Transport, 2023, 36(9): 257-270.
[7] Montanino M, Punzo V. Trajectory data reconstruction and simulation-based validation against macroscopic traffic patterns[J]. Transportation Research Part B: Methodology, 2015, 80: 82-106.
[8] 李依, 赵建有, 叶云涛.NGSIM轨迹重构研究综述[J].汽车实用技术, 2021, 46(4): 204-208.
Li Yi, Zhao Jian-you, Ye Yun-tao, et al.A review of NGSIM trajectory reconstruction[J].Automobile Applied Technology, 2021, 46(4): 204-208.
[9] 刘阳林. 无人驾驶车辆局部换道决策与换道轨迹规划算法研究[D].长春: 吉林大学汽车工程学院, 2023.
Liu Yang-lin. Research on local lane change decision and lane change trajectory planning algorithm of driverless vehicle[D]. Changchun: College of Automotive Engineering, Jilin University, 2023.
[10] 赵志敏. 数据驱动的车辆换道意图识别及轨迹预测研究[D].北京: 北京交通大学交通运输学院, 2022.
Zhao Zhi-min. Data-driven vehicle lane change intention recognition and trajectory prediction[D]. Beijing: School of Traffic and Transportation, Beijing Jiaotong University, 2022.
[1] Yi-yong PAN,Sai-sai YANG. Analysis of correlation between self⁃driving car takeover mode and accident severity [J]. Journal of Jilin University(Engineering and Technology Edition), 2026, 56(8): 2077-2083.
[2] Yun-wei MENG,Zhi-peng LI,Yin ZHANG,Zhen-yu QUAN,Guang-qing YANG,Fang CHEN,Si-jing LAI,Jun QIAO. Driving comfort in tunnel groups based on longitudinal spacing and brightness difference [J]. Journal of Jilin University(Engineering and Technology Edition), 2026, 56(7): 1904-1914.
[3] Jin-peng WANG,Shun-yao DOU,Jue WANG,Xin ZHAO. Optimization algorithm for transmission rate improvement and energy efficiency in MIMO-NOMA system based on a deep learning CCDNN framework [J]. Journal of Jilin University(Engineering and Technology Edition), 2026, 56(3): 830-842.
[4] Yu-zhao ZHANG,Xin-ru DING,Chang-xiao HOU,Hu-lin CHEN. Research status and prospects of resilience for transportation networks [J]. Journal of Jilin University(Engineering and Technology Edition), 2026, 56(3): 585-602.
[5] Dan YUE,Chong-shuai WANG,Ya-ting YANG,Hai-tao NIE,Ya-hui CHUAI. High resolution computational imaging technology for a simple adaptive optics system [J]. Journal of Jilin University(Engineering and Technology Edition), 2026, 56(2): 543-554.
[6] Xiao-dong CAI,Shun-hong LONG,Kun-jun LIANG. Video captioning method based on enhanced object learning and attention networks [J]. Journal of Jilin University(Engineering and Technology Edition), 2026, 56(2): 516-522.
[7] Zhuang-lin MA,Yu-ming BI,Bei ZHOU,Ya-juan DENG,Xue ZHAO. Heterogeneity analysis of residents’ transfer intentions under transit transfer preferential policy [J]. Journal of Jilin University(Engineering and Technology Edition), 2026, 56(1): 158-169.
[8] Zong-wei YAO,Chen CHEN,Zhen-yun GAO,Hong-peng JIN,Hao RONG,Xue-fei LI,Hong-pu HUANG,Qiu-shi BI. Visual recognition of excavator keypoints based on synthetic image datasets [J]. Journal of Jilin University(Engineering and Technology Edition), 2026, 56(1): 76-85.
[9] Lin-hong WANG,Yu-yang LIU,Zi-yu LIU,Ying-jia LU,Yu-heng ZHANG,Gui-shu HUANG. Defect recognition of lightweight bridges based on YOLOv5 [J]. Journal of Jilin University(Engineering and Technology Edition), 2025, 55(9): 2958-2968.
[10] Yun-xiang ZHANG,Xian-min SONG,Yu XIE,Tian-shu ZHAN. Agent⁃based behavioral simulation of parking reservation services driven by user satisfaction [J]. Journal of Jilin University(Engineering and Technology Edition), 2025, 55(9): 2978-2984.
[11] Hui XU,Jia-le XIA,Zhen-yao MA,De-jun BIAN,Sheng-shu AI. Optimization strategy for sewage heat extraction system based on adaptive weight particle swarm [J]. Journal of Jilin University(Engineering and Technology Edition), 2025, 55(9): 3049-3055.
[12] Jing LIAN,Ji-bao ZHANG,Ji-zhao LIU,Jia-jun ZHANG,Zi-long DONG. Text-based guided face image inpainting [J]. Journal of Jilin University(Engineering and Technology Edition), 2025, 55(8): 2732-2740.
[13] Yuan-ning LIU,Xing-zhe WANG,Zi-yu HUANG,Jia-chen ZHANG,Zhen LIU. Stomach cancer survival prediction model based on multimodal data fusion [J]. Journal of Jilin University(Engineering and Technology Edition), 2025, 55(8): 2693-2702.
[14] Shu-shan CHAI,Zhi-qiang ZHOU,Hai-tao LI,Jiong-yang XU. Real-time road network traffic anomaly incident detection based on graph spatial-temporal pattern learning network [J]. Journal of Jilin University(Engineering and Technology Edition), 2025, 55(7): 2145-2161.
[15] Jiang-bo YU,Jian-cheng WENG,Peng-fei LIN,Yu-xing SUN,Jiao-long CHAI. Passenger flow prediction model of external transportation hub based on hybrid Transformer [J]. Journal of Jilin University(Engineering and Technology Edition), 2025, 55(7): 2251-2259.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
Copyright © Journal of Jilin University(Engineering and Technology Edition), All Rights Reserved.
Tel: +86-431-85095297　Fax: +86-431-85094074　E-mail: xbgxb@jlu.edu.cn
Powered by Beijing Magtech Co. Ltd