Research On Safety Risk Assessment Method Of Driver's Illegal Use Of Mobile Phone Based On The Cognition Of Visual Perception

With the popularity of smart phones,mobile phones are closely related to people's lives,which lead to the undesirable phenomena of "cellphones never leave hands" and "hands never leave cellphones".Although the ubiquitous mobile phones provide people with a lot of convenience,they bring troubles as well,such as illegal use of mobile phones during driving.This behavior occupies the driver's visual resources seriously.In recent years,the illegal use of mobile phones in driving has turn into one of the most significant causes of road traffic accidents.At the present stage,there are some problems in the driving behavior and safety research,such as lack of correlation with individual attributes,incomplete mobile phone operation,lack of visual representation parameters,and imperfect safety assessment.To this end,this paper takes the driving safety of illegal use of mobile phone as the research object,based on the survey of using mobile phone while driving and individual correlation analysis,reveals the visual behavior characteristics corresponding to the driver's illegal use of mobile phones,and builds a driving safety evaluation index system.To study the safety risk assessment method for illegal use of mobile phones while driving.The research results have practical significance in standardizing driver behavior,reducing road traffic accidents caused by illegal use of mobile phones and design of safety assisted driving systems.The main research works of this dissertation are follows:Firstly,in order to investigate the current situation mobile phones' use during driving,relevant mobile phone data are collected by questionnaire survey.The Person correlation between the mobile phones usage and the individual attributes of the respondents is used to identify the influencing factors of the driver's use of the mobile phone.Secondly,based on the questionnaire data analysis,select five driving states related to mobile phones,such as driving normally,making calls,sending and receiving messages,watching videos,and updating WeChat moments,and devise a driving experimental scheme for illegal use of mobile phones to obtain visual parameters under various driving conditions.According to the characteristics of visual parameter data,the Pauta criterion is used to clean the anomaly data.Combined with the characteristics of the data,the joint frequency distribution histogram is introduced to lay the foundation for the dynamic division of the driver's gaze area.Then,the visual parameters are initially determined by combing the dynamic visual behavior theory and selecting the classification criteria of fixation and saccade suitable for this experiment.On the basis of data preprocessing,k-means clustering algorithm and hierarchical clustering algorithm are applied to gaze.Through contrastive analysis,optimizing the scheme of dynamic gaze region division;based on this,observes the impact of illegal use of mobile phones on the visual behavior of drivers from four dimensions: fixation,saccade,blink and pupil,and the index system for evaluating driving safety is constructed.Finally,based on the influence of five driving states on visual parameters,the principal component analysis(PCA)method is used to judge the safety level of different driving processes.The artificial neural network(ANN)is utilized to detect different driving states.The ANN and PCA method are used in combination to build a real-time safety risk assessment model of using mobile phones while driving...