基于多通道信息融合的疲劳驾驶行为分析研究  

Research on Drowsy Driving Behavior Based on Multi-channel Information Fusion

作　　者：秦洪懋[1] 刘志强[1] 汪澎[1] 

机构地区：[1]江苏大学汽车与交通工程学院,江苏镇江212013

出　　处：《中国安全科学学报》2011年第2期115-120,共6页China Safety Science Journal

基　　金：国家科技支撑计划课题(2007BAK35B02)

摘　　要：为了克服单一通道信息在驾驶疲劳行为判定中的局限性,提出了综合运用多通道信息融合共同判定驾驶疲劳行为的方法。该方法在充分考虑各信息源相关性和互补性的基础上,优化采用驾驶人疲劳特征ECD、车道偏离程度SAAE、方向盘转动程度SWA等疲劳判别指标,运用MVAR进行多维特征向量提取,以有向无环支持向量机为融合算法,建立了基于多分类支持向量机的驾驶疲劳行为判定模型。结果表明,运用DAG-SVM进行多通道信息决策提高了疲劳驾驶行为检测的准确性和可靠性。In order to overcome the limitations of single-channel information in the determination of drowsy driving behavior,a method was proposed based on multi-channel information fusion.According to the correlation and complementarities analysis,some parameters such as ECD（Eye Closure Degree） of driver fatigue characteristics,SAAE（Symmetry Axis Angle of EDF） of lane deviation,and SWA（Steering Wheel Angle）of steering wheel rotation degree were developed,multivariate autoregressive（MVAR） model was studied to extract multi-dimensional feature vectors,and the determination mode of drowsy driving behavior was established based on the multi-class support vector machine with DAG-SVM（Directed Acyclic Graph-Support Vector Machine）.The results show that the proposed model with DGA-SVM improves the reliability and accuracy of drowsy driving behavior detection.

关 键 词：交通运输安全工程 驾驶辅助系统 疲劳驾驶 支持向量机(SVM) 信息融合 

分 类 号：X913.4[环境科学与工程—安全科学]