Analysis Of Psycho-behavioral Correlates Of Elderly People’s Waiting At Signalized Intersections For Crossing The Street

The elderly population in China is increasing year by year,and the trend of social aging is increasingly prominent.With the development of cities,the traffic environment is becoming increasingly complex,and the number of traffic accidents and casualties among the elderly accounts for a significant proportion.Signalized intersections are the most complex traffic network nodes faced by the elderly.Elderly people have characteristics such as decreased physical function,susceptibility to stress and fatigue,and reduced responsiveness,making them more prone to accidents when crossing the street.In this context,the visual search characteristics and physiological and psychological states of the elderly when waiting for crossing at signal intersections are studied.By establishing the correlation between psychology and behavior,then improve the crossing environment of the elderly and enhance their crossing safety,which is the basic work of traffic aging construction and has important practical significance.In order to investigate the visual search characteristics and physiological and psychological states of elderly people waiting for crossing at signal intersections,this thesis selected and photographed two signal intersection scenes as experimental stimulus materials,carried out visual search and physiological and psychological response experiments of pedestrians waiting for crossing at signal intersections.Gaze,sweep,ECG and electrodermal data were collected with the aid of oculomotor and physiological instruments.A total of 120 sets of valid eye-movement and physiological data were obtained from 30 elderly people and 30 young people in Scenario A and Scenario B,simulating the waiting process of crossing the street.Based on the experimentally obtained data,firstly,the visual search and physiological-psychological response characteristics of the elderly and young people during the street crossing and waiting were compared and analyzed.The results show that there are differences between the elderly and the young people in processing the environmental information at intersections.To ensure their crossing safety,the elderly need longer and repeated gaze behaviors to fully perceive and understand environmental information.The elderly need to obtain more information about the traffic environment than younger people.Intersection environments with varying degrees of complexity may have varying degrees of impact on the visual search and physiological and psychological responses of the elderly and young.Secondly,the differences in cognitive load and observation perception efficiency between the elderly and young people in the two scenarios were compared and evaluated by means of data envelopment analysis.The results show that:when waiting for crossing at signalized intersections,compared with young people,the elderly have lower observation-perception efficiency,higher cognitive load,and more difficulty in perceiving and understanding traffic information.When facing intersection scenarios with more complex traffic environments,the elderly have greater psychological stress and emotional tension.Gaze point,pupil diameter,sweep speed,heart rate growth rate,HF,HF/LF,EAR amp.and EDR rit.can effectively reflect the observation perception efficiency of the elderly.Finally,three machine learning algorithms were used to combine pedestrians’age,gender,observation perception efficiency values,eye movements and physiological parameters to construct a pedestrian crossing decision prediction model.In addition,eye movement and physiological data were collected to simulate the pedestrian crossing and waiting process in scenarios C and D,and effective sample data from 29 elderly and 38young people were finally obtained to validate the crossing decision prediction model.The results show that the three machine learning models,random forest,XGBoost and support vector machine,are all effective in predicting the crossing decision of pedestrians waiting for crossing the street.The accuracy,recall,precision and F₁scores of the model performance indexes of the random forest model are better than the other two models in scenario A and scenario B.The contribution rates of each feature index in the decision classification task are different.The random forest model,total sweep amplitude,observation perception efficiency value,sum of gaze point durations,and number of sweeps have a greater impact on the crossing decision.In the XGBoost model,total sweep amplitude,observation perception efficiency value,sum of gaze point durations,and average pupil diameter of all gaze have a greater impact on the crossing decision.It is verified that in the crossing scenarios of other signal intersections,by building random forest,XGBoost and support vector machine models are used to predict pedestrians’crossing decision by inputting their gaze,sweep,ECG and electrodermal state indicators,as well as their gender,age and observation perception efficiency values.