Open AccessDriver fatigue evaluation model with integration of multi‐indicators based on dynamic Bayesian network
Author(s) -
He Qichang,
Li Wei,
Fan Xiumin,
Fei Zhimin
Publication year - 2015
Publication title -
iet intelligent transport systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.579
H-Index - 45
eISSN - 1751-9578
pISSN - 1751-956X
DOI - 10.1049/iet-its.2014.0103
Subject(s) - dynamic bayesian network , electroencephalography , computer science , artificial neural network , bayesian network , simulation , driving simulator , noise (video) , trunk , artificial intelligence , psychology , psychiatry , image (mathematics) , ecology , biology
Electroencephalogram (EEG) data are an effective indicator to evaluate driver fatigue, but it is usually disturbed by noise. The frequent head nodding, as well as the time of day and total driving time, also have very close relationship with driver fatigue. All these factors should be taken into account for comprehensive driver fatigue evaluation. 50 drivers are recruited to take part in the fatigue‐oriented experiment on the driving simulator. Based on the EEG samples, the EEG‐based indicator of driver fatigue has been established by artificial neural network. Subsequently, a new algorithm is present to compute the head nodding angle with posture data from the passive tools fixed on the driver's head and trunk, respectively, and then head‐based indicator of driver fatigue is determined. Finally, a new evaluation model of driver fatigue is established with integration of four fatigue‐based indicators with DBN (Dynamic Bayesian Network). The results show that it is more accurate to evaluate the driver fatigue compared with the sole EEG‐based indicator.