American Society of Civil Engineers


A Driver Fatigue Recognition Model Based on BP Neural Network


by Shiwu Li, (College of Transportation, Jilin University, Changchun, 130025. E-mail: shiwu@jlu.edu.cn), Linhong Wang, (College of Transportation, Jilin University, Changchun, 130025. E-mail: linhong0520@yahoo.com.cn), Zhifa Yang, (College of Transportation, Jilin University, Changchun, 130025. E-mail: wanglh09@mails.jlu.edu.cn), Jingjing Tian, (College of Transportation, Jilin University, Changchun, 130025. E-mail: 395815626@163.com), and Dongmei Si, (College of Transportation, Jilin University, Changchun, 130025. E-mail: Sidongmei27@163.com)
Section: Traffic Safety and Emergency Management, pp. 2072-2079, (doi:  http://dx.doi.org/10.1061/41186(421)205)

     Access full text
     Purchase Subscription
     Permissions for Reuse  

Document type: Conference Proceeding Paper
Part of: ICCTP 2011: Towards Sustainable Transportation Systems
Abstract: This article presents a driver fatigue recognition model based on a BP neural network that was developed to identify a driver’s fatigue level both subjectively and precisely. Fatigue level was classified into the following four levels: alert, mild fatigue, deep fatigue, and drowsiness. Alpha rhythms, beta rhythms, delta rhythms, the mean of heart rate, and standard deviation of R-R intervals were selected as input parameters of the BP neural network and the driver’s fatigue level was set as the output parameter. Part of the experimental data, collected from the JiangYan freeway, was used to train the BP neural network model, and the parameters of the model were calibrated. Then residual data was used to test the effectiveness of the model. Results showed that the model can recognize driver fatigue levels with high precision.


ASCE Subject Headings:
Driver behavior
Neural networks
Traffic safety