计算机技术与发展2025，Vol.35Issue(3)：133-139,7.DOI:10.20165/j.cnki.ISSN1673-629X.2024.0332
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
摘要
Abstract
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%,whichis12.8 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Key words
fatigue driving/mountain roads/channel attention/sparrow search algorithm/bidirectional gated recurrent unit分类
计算机与自动化引用本文复制引用
李作进,李东阳,蔡俊锋,李明虹,彭大兵,郑路..基于改进BiGRU网络的山地道路疲劳驾驶识别方法[J].计算机技术与发展,2025,35(3):133-139,7.基金项目
重庆市自然科学基金(CSTB2023NSCQ-MSX0760) （CSTB2023NSCQ-MSX0760）
重庆市教委科技重大项目(KJZD-M202301502) （KJZD-M202301502）
