Research On Distracted Driving By Mobile Communication Software Based On Driving Simulator

With the widespread popularity of smart terminals and the Internet,the problem of drivers using mobile phones while driving has become more and more common,and the problem of distracted driving has gradually become an important cause of traffic accidents.Communication software integrates text,voice,and multimedia technologies,making it a cheaper and more convenient way of communication.At the same time,when attention is occupied by other things(such as driving status),people are more inclined to use voice messages than text messages.Research on the effect of voice messages on drivers’ distracted behavior is still lacking.Therefore,this article studies the distracted driving behavior caused by different interaction methods(text messages and voice messages),and builds a discriminant model of distracted driving to provide theoretical support for improving road safety.First,based on the driving simulator and eye tracking equipment,this article designs distracted driving simulation experiments for "receiving text messages" and "receiving voice messages".According to the types of high-incidence accidents in my country,dangerous scenes with complex traffic environments are designed.The dangerous scenes include four types of dangerous interferences: pedestrian crossing,vehicle lane change,front vehicle braking,and intersection conflict,which have good simulation effects and universal applicability.After that,under the three conditions of normal driving,sending and receiving text messages and sending and receiving voice messages,a large number of human driving simulation experiments were carried out using the designed experimental plan,and the data collected by the driving simulator and the eye tracker were integrated.Secondly,based on the data obtained from the driving simulation experiment,the single-factor variance method is used to analyze the impact of different message types on driving safety.Aiming at the interference of two different types of We Chat messages(text messages and voice messages)on the driver’s driving behavior,comparing the driver’s reaction time and braking distance in an emergency situation,it was found that compared with text messages,voice messages were sent and received.Information has the same degree of negative influence on the driver’s distracted behavior,so it proves the necessity of the voice message to be included in the recognition of distracted driving behavior.Analyze the vehicle motion state and the driver’s visual characteristics under normal and information interference conditions,obtain a set of indicators related to distracted driving behavior,and find that the vehicle motion state and the driver’s visual characteristics under two types of messages are affected The degree is different,which further proves the necessity of the voice message to be included in the recognition of distracted driving behavior.Finally,the level of driving distraction is divided into three categories,and sample data of different distraction levels are extracted.Based on the results of significant differences,a set of indicators for discriminating distracted driving is constructed.Based on support vector machine and genetic algorithm,a GA-SVM discriminating model of distracted driving is constructed.The accuracy and sensitivity are selected as the discriminant indicators to test the effectiveness of the discriminant model when inputting different data types.The results show that compared with a single type of data,the discriminant effect and practicability of the discriminant model fused with multi-source data are greatly improved.