基于改进BiGRU网络的山地道路疲劳驾驶识别方法  

Fatigue Recognition of Drivers on Mountain Roads Based on Improved BiGRU Network

作　　者：李作进 李东阳 蔡俊锋 李明虹 彭大兵 郑路 LI Zuo-jin;LI Dong-yang;CAI Jun-feng;LI Ming-hong;PENG Da-bing;ZHENG Lu(School of Electronic and Electrical Engineering,Chongqing University of Science and Technology,Chongqing 401331,China)

机构地区：[1]重庆科技大学电子与电气工程学院,重庆401331

出　　处：《计算机技术与发展》2025年第3期133-139,共7页Computer Technology and Development

基　　金：重庆市自然科学基金(CSTB2023NSCQ-MSX0760);重庆市教委科技重大项目(KJZD-M202301502)。

摘　　要：针对复杂山地路况下疲劳驾驶行为隐匿性高和识别准确率低的问题,提出一种改进双向门控循环单元(Bidirectional Gated Recurrent Unit,BiGRU)神经网络的山地道路驾驶人疲劳状态识别方法。该方法首先利用BiGRU网络深入挖掘驾驶行为数据中的复杂前后依赖关系,然后引入麻雀搜索算法(Sparrow Search Algorithm,SSA)对BiGRU模型的初始学习率、L2正则化系数、隐藏层神经单元数和最大迭代次数进行优化,以提升模型的收敛速度、泛化能力和稳定性。为了进一步增强模型对疲劳特征的识别能力,在BiGRU模型中嵌入了通道注意力(Channel Attention,CA)机制,通过自适应调整不同特征通道的权重,突出疲劳驾驶与正常操作的差异性,有效提高了识别准确率。实车数据验证表明,该方法在三级疲劳检测中的识别准确率达到92.0%,较原BiGRU模型提升12.8百分点,较其他常用疲劳驾驶检测模型也表现出更好的性能,满足了实际工程需求。To address the challenges of high stealthiness in fatigue driving behavior and low recognition accuracy under complex mountain road conditions,we propose an improved method for recognizing driver fatigue states using an enhanced Bidirectional Gated Recurrent Unit(BiGRU)neural network.This method first utilizes the BiGRU network to deeply explore the complex before and after dependencies in driving behavior data,and then introduces the Sparrow Search Algorithm(SSA)to optimize the initial learning rate,L2 regularization coefficient,number of hidden layer neural units,and maximum iteration times of the BiGRU model,in order to improve the convergence speed,generalization ability,and stability of the model.To further enhance the model’s fatigue recognition capabilities,we integrate the Channel Attention(CA)mechanism into the BiGRU model.This mechanism adaptively adjusts the weights of different feature channels,thereby distinguishing between fatigue driving and normal operation,and significantly improving recognition accuracy.The validation of real vehicle data shows that the recognition accuracy of the proposed method in three-level fatigue detection reaches 92.0%,which is 12.8 percentage points higher than that of the original BiGRU model,and also shows better performance than other commonly used fatigue driving detection models to meet the actual engineering requirements.

关 键 词：疲劳驾驶 山地道路 通道注意力 麻雀搜索算法 双向门控循环单元 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]