融合递归图的脑电驾驶行为分类方法研究

doi:10.16097/j.cnki.1009-6744.2025.03.014

摘要/Abstract

摘要： 驾驶行为识别是智能驾驶辅助系统的核心问题，而基于脑电信号的驾驶行为分类检测是实现以人为中心的智能驾驶辅助系统的重要途径。为实现常见驾驶行为下脑电信号五分类检测，本文提出一种基于递归图和改进卷积神经网络的方法，即RP-CS。RP-CS方法主要通过挖掘驾驶员脑电信号中重要的非线性特征，并通过融入了通道注意力机制的卷积神经网络进一步提取特征并实现五分类检测。在本方法中，脑电非线性特征的提取方式是将一维时序信号嵌入到更高维度的相空间中，并利用欧氏距离构造出递归图。随后，将同时含有非线性特征和时域特征的递归图作为改进卷积神经网络的输入，利用卷积神经网络在图像数据处理中的优势，实现对驾驶行为的五分类检测。在公开数据集上进行对比实验，结果表明，本文提出的RP-CS驾驶行为识别准确率优于其他方法。在单被试条件下，五类驾驶行为的分类准确率最高可达95.84%；在跨被试条件下，平均分类准确率为71.92%。此结果意味着脑电信号中非线性特征和深度学习模型的有效结合可提高脑电信号分类识别的效率。为驾驶行为监测和辅助安全驾驶提供了一种可行的解决方案，对于提高驾驶辅助系统的性能和驾驶安全具有重要意义。

关键词: 智能交通, 驾驶行为分类, 深度学习, 脑电, 非线性特征

Abstract: Driving behavior recognition is a core challenge in intelligent driving assistance systems, and classifying driving behaviors based on Electroencephalography (EEG) signals is crucial for achieving human-centered intelligent driving assistance. To enable five-class classification of EEG signals under common driving behaviors, this paper proposes a method based on recurrence plots and a convolutional neural network (CNN) with channel squeeze enhancement (RP-CS). The RP-CS method extracts nonlinear features from EEG signals by embedding one-dimensional time-series signals into a higher-dimensional phase space and constructing recurrence plots using Euclidean distances. These recurrence plots, which integrate both nonlinear and temporal features, are used as input to a CNN enhanced with a channel attention mechanism, enabling accurate five-class classification. Comparative experiments on a public dataset demonstrate that the RP-CS method achieves superior performance, with a maximum classification accuracy of 95.84% under subject-dependent conditions and an average classification accuracy of 71.92% under subject-independent conditions. The results indicate that effectively combining nonlinear EEG features with deep learning models can significantly enhance the efficiency of EEG-based classification. This approach offers a viable solution for driving behavior monitoring and safety assistance, contributing to improved performance in driving assistance systems and enhanced driving safety.

Key words: intelligent transportation, driving behavior classification, deep learning, electroencephalogram (EEG), nonlinear characteristics

中图分类号:

TP391
TP18

常文文, 芦家磊, 黄霄, 闫光辉. 融合递归图的脑电驾驶行为分类方法研究[J]. 交通运输系统工程与信息, 2025, 25(3): 152-162.

CHANG Wenwen, LU Jialei, HUANG Xiao, YAN Guanghui. Driving Behavior Classification Using Electroencephalogram and Recurrence Plot[J]. Journal of Transportation Systems Engineering and Information Technology, 2025, 25(3): 152-162.

导出引用管理器 EndNote|Ris|BibTeX

链接本文: http://www.tseit.org.cn/CN/10.16097/j.cnki.1009-6744.2025.03.014

http://www.tseit.org.cn/CN/Y2025/V25/I3/152

参考文献

[1]World Health Organization. Global status report on road safety 2018[M]. Geneva: World Health Organization, 2019.

[2]HAGHANI M, BLIEMER M C J, FAROOQ B, et al. Applications of brain imaging methods in driving behaviour research[J]. Accident Analysis & Prevention, 2021, 154: 106093.

[3]ZHAO J, ZHAO W, DENG B, et al. Autonomous driving system: A comprehensive survey[J]. Expert Systems with Applications, 2024, 242: 122836.

[4]SONKO S, ETUKUDOH E A, IBEKWE K I, et al. A comprehensive review of embedded systems in autonomous vehicles: Trends, challenges, and future directions[J]. World Journal of Advanced Research and Reviews, 2024, 21(1): 2009-2020.

[5]BENTO V, PAULA L, FERREIRA A, et al. Advances in EEG-based brain-computer interfaces for control and biometry[C]//International Symposium on Computational Intelligence for Engineering Systems, 2009: 1-14.

[6]NACPIL E J C, WANG Z, GUAN M, et al. EEG-based emergency braking prediction using data ablation and SVM classification[J]. IEEE Sensors Journal, 2023, 23 (14): 16013-16019.

[7]HU J, MIN J. Automated detection of driver fatigue based on EEG signals using gradient boosting decision tree model[J]. Cognitive Neurodynamics, 2018, 12: 431-440.

[8]AREFNEZHAD S, SAMIEE S, EICHBERGER A, et al. Applying deep neural networks for multi-level classification of driver drowsiness using vehicle based measures[J]. Expert Systems with Applications, 2020, 162: 113778.

[9]张冰涛,常文文,李秀兰.基于时空脑电特征与并行神经网络的疲劳驾驶检测[J].交通运输系统工程与信息, 2023, 23(2): 315-325. [ZHANG B T, CHANG W W, LI X L. Fatigue driving detection based on spatiotemporal EEG features and parallel neural networks [J]. Journal of Transportation Systems Engineering and Information Technology, 2023, 23(2): 315–325.]

[10] LI J, LIANG X, LIU Y, et al. Decoding braking intentions from EEG signals: A convolution-transformer network approach[J]. IEEE Sensors Journal, 2024, 24(22): 38329-38343.

[11] LIANG X, YU Y, LIU Y, et al. EEG-based emergency braking intention detection during simulated driving[J]. BioMedical Engineering OnLine, 2023, 22(1): 65.

[12] JU J, BI L, FELEKE A G. Noninvasive neural signal based detection of soft and emergency braking intentions of drivers[J]. Biomedical Signal Processing and Control, 2022, 72: 103330.

[13] LI M, WANG W, LIU Z, et al. Driver behavior and intention recognition based on wavelet denoising and Bayesian theory[J]. Sustainability, 2022, 14(11): 6901.

[14] TAO X, GAO D, ZHANG W, et al. A multimodal physiological dataset for driving behaviour analysis[J]. Scientific Data, 2024, 11(1): 378.

[15] YANG L, MA R, ZHANG H M, et al. Driving behavior recognition using EEG data from a simulated car following experiment[J]. Accident Analysis & Prevention, 2018, 116: 30-40.

[16] WANG Y, LIU T, QIN Y, et al. A hybrid approach for driving behavior recognition: Integration of CNN and transformer-encoder with EEG data[C]//2023 IEEE 98th Vehicular Technology IEEE, 2023: 1-5.

[17] 赵朔, 奇格奇,李培豪,等.基于脑电通道注意力机制的驾驶行为识别研究[J].交通运输系统工程与信息,2024, 24(4): 283-291. [ZHAO S, QI G Q, LI P H, et al. Driving behavior recognition based on EEG channel attention mechanism[J]. Journal of Transportation Systems Engineering and Information Technology, 2024, 24(4): 283-291]

[18] STAM C J. Nonlinear dynamical analysis of EEG and MEG: Review of an emerging field[J]. Clinical Neurophysiology, 2005, 116(10): 2266-2301.

[19] 冯国红,郑潇,张彬,等.基于独立成分分析:递归图和改进的高效能网络的脑电情绪识别研究[J].生物医学工程学杂志,2024,41(6): 1103-1109. [FENG G H, ZHENG X, ZHANG B, et al. EEG emotion recognition based on independent component analysis: Recurrence plot and improved efficientNet[J]. Journal of Biomedical Engineering, 2024, 41(6): 1103-1109.]

[20] HU J, SHEN L, SUN G. Squeeze-and-excitation networks [C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018: 7132-7141.

[21] THIEL M, ROMANO M C, KURTHS J. How much information is contained in a recurrence plot?[J]. Physics Letters A, 2004, 330(5): 343-349.

[22] MARWAN N, ROMANO M C, THIEL M, et al. Recurrence plots for the analysis of complex systems[J].Physics Reports, 2007, 438(5/6): 237-329.

[23] BRAUN T, KRAEMER K H, MARWAN N. Recurrence flow measure of nonlinear dependence[J]. The European Physical Journal Special Topics, 2023, 232(1): 57-67.

[24] MARWAN N, ROMANO M C, THIEL M, et al. Recurrence plots for the analysis of complex systems[J]. Physics Reports, 2007, 438: 237-329.

[25] LIN C T, KING J T, CHUANG C H, et al. Exploring the brain responses to driving fatigue through simultaneous EEG and fNIRS measurements[J]. International Journal of Neural Systems, 2020, 30(1): 1950018.

[26] 袁月婷, 闫光辉,常文文,等.基于脑电信号空域特征的紧急制动行为识别[J].电子科技大学学报,2024,53(1): 84-91. [YUAN Y T, YAN G H, CHANG W W, et al. Emergency braking behavior recognition based on spatial features of EEG signals [J]. Journal of University of Electronic Science and Technology of China, 2024, 53 (1): 84-91.]

[27] LAWHERN V J, SOLON A J, WAYTOWICH N R, et al. EEGNet: A compact convolutional neural network for EEG-based brain-computer interfaces[J]. Journal of Neural Engineering, 2018, 15(5): 056013.

[28] SHI C, LIU L, ZHANG C, et al. To investigate the ability of CNN in learning specific frequency band of motor imagery EEG[C]//2022 7th International Conference on Signal and Image Processing (ICSIP), IEEE, 2022: 650-654.