Research On Urban Rail Transit Passenger Flow Analysis Platform Based On Spark

At present,the urban rail transit of China is in a stage of rapid development,with the length of lines,the number of locomotives and passenger capacity growing year by year.Large-scale expansion is accompanied by the growth of urban rail transit’s data,which brings the demand of data storage,management and analysis.In order to improve the efficiency of daily operation,passenger organization,and further to enhance the level of service,the demand for the information of real-time passenger flow state is getting intense,while the traditional data processing methods have been unable to meet.Thus,a real-time passenger flow data processing method is needed to reflect the change of passenger demand and the whole state of the system in time,providing convincing decision support for operational organization and planning adjustment.This paper,focusing on the requirements of monitoring passenger flow status and statistical analysis of dynamic data and in order to solve the problems in data processing and application,analyzed the technical characteristics,advantages,limitations and applicable environment of the current mainstream data processing framework.Based on the current content,scale,characteristics and application status of urban rail transit passenger flow data,this paper discussed the necessity of research on passenger flow analysis platform with big-data processing ability.Firstly,the functional and non-functional requirements of the passenger flow analysis platform based on big-data processing technology are analyzed,then the computer cluster deployment and the architecture of platform based on the hybrid processing framework Spark are designed,which include data acquisition layer,data storage and management layer,computing processing layer and application layer.Secondly,taking Beijing urban rail transit system as the research object,we analyzed the passenger flow statistical characteristics under different time grading based on the static data(such as information of network,station and operation plan)and dynamic passenger flow data.Combining the theory of dynamic OD matrix estimation,passenger flow distribution and prediction algorithm,we designed the data processing and analysis algorithm,completing the function design of online traffic status monitoring,trend forecast for the platform.Finally,to realize the visual expression of data results,an urban rail transit network visualization module was constructed according to the research object,and a platform cluster operating environment was built for simulation test.The test results shows that the proposed design and technical methods are feasible and effective.The urban rail transit passenger flow analysis platform,with the ability of big-data analysis and processing,is able to provide dynamic basis and decision support for daily operation management,passenger organization,which is helpful to promote the operation efficiency and service level.