Mechanism Analysis Of Pedestrian Distracted By Smartphone While Street Crossing

With the rapid development of electronic devices such as mobile phones,the number of mobile phone users increases dramatically in recent years.However,accompanied with convenience,the distracting behavior caused by over-reliance on smart devices also brings potential risks to people’s lives.Among them,the distracted walking behavior of using mobile phones while crossing the street will increasingly pose unaware dangers to pedestrians.Therefore,this paper aims at the characters of distracted walking behaviors that pedestrians using mobile phones when crossing the street,and explores the mechanism of behaviors.First,in order to explore the behavior characteristics of people using mobile phones while crossing the s treet,a video shooting method was used to observe the behavior of 2515 pedestrians at eight intersections in Baise,of which 10.19% pedestrians used smartphone while acrossing the street.The analysis of behavior characteristics shows that all kinds of smartphone usage will occupy people’s attention resources to varying degrees.A Poisson regression model was established to explore how the environmental and road factors will affect the using rate of smartphone while crossing the street,the results show that traffic volume will reduce the using rate of mobile phones,while signalized intersections,number of lanes,road greening rate will increase the phone using rate significantly.Secondly,based on the extended planned behavior theory,a questionnaire and the hypotheses about the mechanism of distracted walking are put forward.322 valid questionnaires are collected through the network.Further,through exploratory factor analysis,a total of 9 distracted walking-related factors were extracted,and a structured equation model was used to verify and analyze the extended planned behavior theory.The results show that 6 factors including subjective attitude,mobile phone dependence,etc.significantly influence on behavioral intention and behavior frequency.Factor scores were obtained based on the imputation theory,and ANOVA was used to analyze the differences of factors scores between groups.A total of 8 groups including age,whether they are students,and mobile phone usage habits,etc.are identified to be significantly different.Finally,based on the verified structural equation model and the ANOVA,a Bayesian network inference model of distracted walking behavior was established.Through structure learning and parameter learning,the association hierarchical relationship of distracted walking is obtained,and the validity of the model is verified by comparing the conditional probability of nodes and actual samples.At the same time,by using indicators such as accuracy,sensitivity,and ROC curve,it shows that the model fit well.Using Bayesian network joint tree inference engine and backward inference to further analyze the probability distribution of the connections between the various levels under the influence of different factors to obtain the distracted walking behavior of pedestrians.Taken together,these may serve as a useful guide for developing legislation and technical countermeasures to ensure pedestrians safety by reducing the distracted walking when crossing the road.