基于驾驶模拟实验的眼部指标与疲劳分级

Eye Indicators and Drowsy Level Analysis Based on Driving Simulator

摘要:

基于同济大学8自由度高仿真驾驶模拟器，利用眼动仪采集了15位驾驶员各一小时驾驶过程中眼部指标，将驾驶员的疲劳状态分为清醒、轻度疲劳、中度疲劳、重度疲劳和极度疲劳5个等级，建立了考虑驾驶员个体差异的分层有序离散选择模型.结果表明：分层有序离散选择模型能有效地考虑驾驶员的个体差异，提高了对驾驶员不同疲劳等级的识别准确率.

关键词:驾驶疲劳;驾驶模拟器;分层有序离散选择模型;闭眼百分率(PERCLOS);瞳孔直径变异系数;眨眼时间

Abstract:

Based on Tongji 8 degrees of freedom driving simulator, 15 samples’ eye indicators were collected using the SmartEye. The driver’s fatigue was divided into five levels, including awake, mild fatigue, moderate fatigue, severe fatigue and extreme fatigue. Taking into account of the drivers’ inter individual differences, a multilevel ordered logistic model was applied. The results show that the multilevel ordered logistic model can consider the inter individual differences among drivers effectively and improve the recognition accuracy of drowsy levels.

Key words:driver fatigue; driving simulator; multilevel ordered logistic model; percentage of eyelid closure over the pupil over time(PERCLOS); variable coefficient of pupil diameter; blink duration