论文部分内容阅读
研究表明疲劳驾驶是引起交通事故的重要原因之一,因此有必要采取预防措施,而能够提前对事故进行准确预报并保证低误报警率是问题的关键所在.提出了利用多眼睑运动特征参数建立支持向量机模型进行疲劳预测的方法,其中眼睑运动特征参数是从驾驶模拟器上采集的眼电信号提取出的.根据Karolinska睡眠等级选出25名缺乏睡眠并在实验中撞到振动带的驾驶员,保证其开始驾驶阶段是警觉的,而事故发生阶段是疲劳的,然后将20名驾驶员作为训练对象,另5名驾驶员作为验证对象.结果表明,所用的方法可以提前至少5 min对由疲劳导致的事故进行预报.
Research shows that fatigue driving is one of the important causes of traffic accidents, so it is necessary to take precautionary measures, and to predict accurately the accident ahead of time and ensure the low false alarm rate is the key issue.The paper proposes the use of multi-eyelid motion parameters Support vector machine model fatigue prediction method, in which eyelid motion parameters were collected from the driver simulator ocular signals extracted.According to Karolinska sleep grade selected 25 sleepless and hit the vibration band in the experiment driving 20 drivers were taken as training subjects and the other 5 pilots as verification subjects.The results showed that the method used could be at least 5 min ahead of time Accidents caused by fatigue are forecast.