Research On The Characteristics Analysis And Passenger Flow Prediction Of Bus Travel For The Elderly In Plateau Cities

At present,China’s aging population has entered a period of accelerated scale,and the contradiction between supply and demand of public transport,as one of the main travel modes for the elderly,is increasingly prominent.Especially in economically underdeveloped plateau cities,with the rapid change of the lifestyle of the elderly population leading to the diversification of travel demand,it means that the requirements for public transportation service level will become higher and higher,and the situation that the public transport travel needs of the elderly population cannot be effectively met will be more severe.Analyzing the characteristics of bus travel for the elderly,positioning their main travel needs,and conducting research on future passenger flow prediction are of great significance in providing accurate bus travel services for the elderly and improving their travel quality.Taking Lhasa City with typical plateau city characteristics as an example,this paper analyzes the bus travel characteristics and passenger flow prediction of the elderly population in plateau cities.The characteristics of urban elderly public transport are affected by urban planning layout,traffic supply and environmental changes,and have obvious regularity.Analyzing the characteristics of bus travel for the elderly and conducting targeted passenger flow prediction research can provide scientific guidance for cities to promote the construction of elderly public transport service system.Firstly,the effective bus travel data of the elderly population is sorted out and combined with the existing travel characteristics analysis methods,and the unique climatic environment and urban structure of plateau cities are synthesized,and the research method of bus travel characteristics of the elderly population in plateau cities is extracted.Secondly,the time,spatial and seasonal effects of bus travel in the elderly population were analyzed.Finally,based on the analysis of bus travel characteristics of the elderly population,a linear and nonlinear passenger flow prediction model is constructed,and the short-term prediction research of bus passenger flow is carried out.The main research contents of this paper are as follows:(1)To study the characteristics of bus travel in the elderly population in plateau cities.The bus data and processing methods are briefly introduced,the data processing process is described in detail,and the short time series clustering method is used to calibrate the daily bus travel time division of the elderly.The bus travel data of the elderly were collected and visualized by date,week,day,hour,etc.,and the periodic and long-term trend analysis was carried out by HP(Hodrick and Prescott)and BP(band-pass)filtering methods,and the results showed that the number of trips of the elderly in Lhasa decreased significantly on weekends and holidays.There are three peak hours for daily bus travel for the elderly population: 7:00-7:30 am,9:00-11:30 am,and 16:30-18:00 pm;with the warmer weather in Lhasa from May to September,the number of trips made by the elderly in the morning has increased significantly,and the end time of evening travel has been delayed compared with other months.The bus travel of the elderly group is cyclical and the cycle is 7 days,and the number of trips shows a slow upward trend with time.(2)Quantitatively analyze the spatial characteristics of bus travel in the elderly population in plateau cities.Calculating the passenger flow sharing rate of the elderly on bus lines,it is found that the bus sharing rate of each line in Lhasa can be divided into three levels: low,moderate and high.Component analysis was used to quantitatively analyze the spatial influence of bus routes on the travel of the elderly,and the results showed that the travel of the elderly was closely related to the distribution of bus routes and the area reached.Combined with static data such as population distribution,bus route and station distribution,the Arc GIS visualization function is used to qualitatively analyze the spatial characteristics of the elderly population,and it is concluded that the elderly travel mostly for the purpose of work,fitness and shopping.Lines with large passenger flow mainly connect population agglomeration areas and commercial developed economic circles.(3)The short-term passenger flow forecast and long-term passenger flow prediction are studied.Short-term passenger flow forecasting: The short-time series clustering seasonal differential autoregressive sliding average prediction model(STSC-SARIMA)and the long-term short-term memory neural network prediction model(STSC-LSTM)were constructed to predict short-term passenger flow.Long-term passenger flow prediction: Seasonal differential autoregressive sliding average prediction model(SARIMA)and convolutional neural network(CNN)prediction model are established to predict the long-term bus flow of the elderly.Compared with the prediction results,it is concluded that the prediction accuracy of STSC-SARIMA with parameter prediction is higher in the prediction of short-term passenger flow of bus travel in the elderly population.In terms of long-term passenger flow prediction,the non-parameter prediction model(CNN)has higher prediction accuracy.