Study On Drivers' Mind Wandering,mental Workload And Safety Risks In Highway Tunnles

Compared with general highway sections,highway tunnels are more likely to cause traffic crashes due to their large difference in internal and external environment,narrow space and closed environment.Therefore,their operation safety has become the focus of social attention.From the perspective of crash mechanism,driver's risk behavior is the main and direct cause of tunnel traffic crashes,and the factors that determine risk behavior include the driver's individual risk attributes and risky driving states.Only by grasping the influencing factors of risky driving states and the internal mechanism of driving behavior changes,and guiding them in a targeted manner,can we obtain stable and effective risk control effects and improve tunnel driving safety.The main content of the risky driving states are the driver's psychological and physical states.Internal cognitive distraction(Mind wandering)and inappropriate mental workload are the most common and highly risky driving states,under the premise of ignoring special circumstances such as fatigue,drunk driving,road rage,and external distraction.The purpose of this study is to investigate the characteristics of mental workload and mind wandering in different highway tunnel sections,and their influences on driving behavior and driving safety.Simultaneously,the influence of individual risk attributes of drivers(individual differences in mind wandering proficiency and individual differences in driving environment familiarity)on the results of the study was also considered.The on-road driving experiment was carried out in typical short,medium and extra-long highway tunnels.Drivers' eye movements,electrocardiogram and speed data were collected by using eye tracker,dynamic multi-parameter physiological detector and On-Board Diagnostic equipment when they drove through different tunnels.Qualitative and quantitative methods were used to analyze the effects of driving experience(the familiarity with the driving environment),tunnel type and different tunnel sections on eye movement indicators.The characteristics of attention distribution and attention transfer characteristics of drivers in different tunnel environments were investigated to further explore the differences of drivers' visual perception ability in different tunnel sections.Simultaneously,the driving simulation experiment was conducted while the mind-wandering frequency,eye movement and driving behavior data were collected.The effects of different tunnel environments on drivers' mind wandering frequency,visual perception and driving behavior were analyzed.On the basis of the non-intrusive driving behavior indicators,a hybrid deep learning model(PCA-LSTM model)for mind-wandering state discrimination was proposed.Then exploratory factor analysis(EFA)method was used to establish quantitative models among eye movement indicators and between eye movement indicators and drivers' mental workload.Then mental workload of drivers while driving in the different tunnel sections was calculated.The effectiveness and accuracy of the model were verified by NASA-TLX subjective task load scale and heart rate index.Simultaneously,the variation characteristics of driving behavior and the internal psychological mechanism of behavior changes were investigated.On this basis,a quantitative model of driving behavior risk in tunnel sections based on behavioral indicators was established.Then the SOM neural network algorithm were utilized to identify the risk attributes and corresponding risk behaviors in different tunnel sections.Finally,the characteristics of mind wandering,mental workload and driving behavior were used to represent the risk levels of perception,cognition and operation.The Entropy-weighting TOPSIS method was adopted to quantify the comprehensive driving risks of different tunnel sections,and the risk degree of each section was ranked.The results show that in the process of driving,drivers who are familiar with the tunnel driving environment are more efficient in searching and processing information,and they tend to pay more attention to the information from the distance of the road,and had a better driving foresight.With the increase of tunnel length,the search efficiency of information decreases,and the mental workload increases.Drivers' mental workload in different driving sections of a same tunnel was ordered by the entrance section,the exit section and the middle section from high to low.The longer the tunnel length,the greater the influence of the light and dark adaptions at the exit and entrance on the mental workload.Drivers who are familiar with the driving environment have a lower mental workload,and it is quicker for them to make physiological response to the changing environment and return to normal state.Different tunnel driving environments,individual differences of mind-wandering propensity and task proficiency have significant effects on the drivers' mind-wandering frequency,visual perception and driving behavior.The comprehensive risk in different driving sections of short,medium and extra-long tunnels is ordered by the entrance section of an extra-long tunnel,the exit section of an extra-long tunnel,the entrance section of a medium tunnel,the middle section of an extra-long tunnel,the exit section of a medium tunnel,the middle section of a medium tunnel,the entrance section of a short tunnel,the exit section of a short tunnel and the middle section of a short tunnel from high to low.The research results can provide a theoretical basis for effectively improving the level of highway tunnel traffic safety.