Research On Overtaking Behavior And Stress

Fatal accident occurred on these roads,and 35-50% of the driver fatalities related to overtaking behavior.Overtaking is a kind of complex driving behavior with a high mental load.It has high requirements on driver's driving ability.With overtaking motivation,drivers often have driving behavior with high risk,such as speeding,reverse driving and so on.Drivers always perform overtaking behavior under stress,it significantly affects driver's driving performance.In order to improve road safety,it is necessary to conduct research on driver overtaking behavior and subjective pressure.Although this issue has great significance,but research directed to driving behavior is still very lacking.Since the 1990 s,research on road traffic safety from the perspective of ergonomics,the current study is not comprehensive and in-depth.Therefore,it is necessary to explore driving behavior characteristics of drivers with stress from the perspective of human factors engineering,it is important to reduce road accident rates in the community,looking for scientific and effective interventions to promote rapid and stable growth of their ability,forming system appropriate training strategies,improve our system of road traffic management intervention.Based on road traffic environment of China,analyze and refine the typical overtaking scene with stress difference of.Based on Vega,C ++,develop overtaking behavior scene,build up the driving simulation platform for driving behavior and driving performance analysis;and invite appropriate driver sample to obtain experiment data,combined with overtaking intent window determining method,to establish drivers sample library in overtaking intent stage and lane keep phase,providing data support for study.According to DBI scale used in the study of Gulian et al,combined with China's actual traffic characteristics,delete,or modify the scale questions that do not meet China's national conditions,to give preliminary scale DBI-oriented country.Obtain a preliminary DBI scale for China.Data through factor analysis collected from 172 questionnaires divide the new five-factor structure,provide structural data for the driver base stress score calculation.NASA-TLX scale scores were tested as the driving pressure level of experience simulation sections.Monitoring driver's ECG,skin electrical signal,a significant difference was found low-frequency electric power,transcutaneous electrical level at different pressure levels,and the change was begun before the driver perform driving behavior.Combine Variation of physiological signals,establish driver stress evaluation model to predict driving pressure.The difference of eye movements and driving behavior performance between intent stage and lane keeping stage was studied,and the driver performance was compared at different stress simulation sections.Compared with the intent phase,the driver's attention has shifted between different gaze regions.Left rearview concern of driver has increased significantly at overtaking intent stage;the simulation results in all sections support this conclusion.Meanwhile the drivers observed and confirmed the surrounding traffic conditions,panning angle and panning angular velocity increased significantly.Almost no difference between the lateral and vertical movement parameters of the vehicle,only differences in steering angle entropy was found.Obtain the characteristic parameters as an indication of its intention to overtake.analysis variation before and after overtaking process of Physiological signal,eye movement signal,vehicle behavior parameters,according to the stability,diversity and difficulty of obtaining,transcutaneous electrical level,the left mirror gaze number,scan amplitude,scanning speed,steering wheel corner entropy five indicators are selected as identification parameters.Use hidden Markov model theory to establish lane keeping intention recognition model and overtaking intention recognition model.Select part of the sample data with integrity,use Baum-Welch algorithm to calculate the model parameters.The remaining part of the sample data with integrity,on behalf of the two models to identify verified the validity of the model.