R.J. Wang, A.D. Yi

Pages: 139-148

Abstract
In order to overcome the problems of low recognition accuracy and poor recognition effect in traditional methods, a recognition method of visual physiological characteristics of road traffic safety driving behavior is proposed. Firstly, the parallax localization equation is used to locate the driver's eye movement state; Secondly, the smart eye Pro eye tracker is used to collect and identify the driver's head motion parameters, and obtain the position parameters and head rotation parameters; Then, eye movement feature points are extracted by obtaining the optimal foraging path of ant colony and combined with the optimal solution retention strategy; Finally, the driver's head motion parameters, eye movement characteristics and driver's gaze time characteristics are obtained to realize the recognition of visual physiological characteristics of driving behavior. The experimental results show that the highest recognition accuracy of this method is 97%, and the effect of feature point recognition is good.
Keywords: driving behavior; visual physiological characteristics; eye movement state; tdoa location; ant colony