点击次数：

论文类型：期刊论文

发表时间：2017-07-01

发表刊物：SIMULATION-TRANSACTIONS OF THE SOCIETY FOR MODELING AND SIMULATION INTERNATIONAL

收录刊物：SCIE、EI、Scopus

卷号：93

期号：7,SI

页面范围：553-565

ISSN号：0037-5497

关键字：Driver fatigue; Bayesian network; maximum likelihood estimation; junction tree

摘要：Driver fatigue is the major reason for severe traffic accidents. At present, the driver's driving state evaluation, based on multi-source information fusion, has become a hotspot in the research field of vehicle safety assistant driving. The purpose of this paper is to build a Bayesian network model for driver fatigue causation analysis considering several visual cues, such as Percentage of Eyelid Closure over the Pupil over Time, Average Eye Closure Speed, etc. The proposed method was divided into three stages, that is, variables analysis, model structure design, and model parameter determination. Finally, the presented model and algorithm were illustrated with a simulation experiment and conclusions were inferred from the experiment data analysis.