疲劳驾驶状态数据融合技术研究

Research of Data Fusion of Fatigue Driving State

疲劳驾驶已成为造成交通事故的主要原因之一,研究有效检测驾驶员疲劳的方法已刻不容缓。现研究集中在单一信号的采集,容易出现误操作,本文基于两个科研项目和七大专利进行多信号融合的研究与设计。用Kalman滤波算法、Welch算法和Adaboost与CamShift相结合的算法分别来分析通过车速、加速度、侧向位移、脉搏和CCD图像传感器采集进来的信号。最后应用D-S证据理论实行二级信息融合,综合判断驾驶员的疲劳程度并作出相应的措施控制机动车。

Fatigue driving has become one of the main causes of road accidents,so it is urgently needed to reseach an effective method of testing pilot fatigue.The current research is focusing on the collection of single signals,which is prone to error operating.In present paper the research and design of multi-signal fusion were carried out on the basis of two scientific research projects and seven patents.The signals,collected by the sensors for speed,acceleration,lateral displacement,pulse and CCD image,were analyzed through Kalman filtering,Welch,and Adaboost-CamShift algorithms.Finally,D-S evidence theory was applied to realize secondary information fusion,judge pilot fatigue comprehensivly and make corresponding measures to control vehicle.