Research And Application Of Precise Guidance Method For Rail Transit Passengers

With the increase in people’s travel activities and the traffic demand,rail transit has begun to develop rapidly,the scale of the road network has increased,and the passenger flow has become increasingly crowded.In order to alleviate passenger flow congestion,in addition to passenger flow control methods,passenger flow guidance has become a new way and means.However,the current effect of passenger flow guidance in the field of rail transit is not good,the scene is single,and there is a lack of personalized and precise guidance for passengers.Therefore,how to provide passengers with multi-scenario and personalized precise guidance services has become a difficult problem that needs to be studied and solved in rail transit passenger information service and operation management.Based on the above background,this paper studies passenger travel preferences by mining rail transit operation big data,establishes a personalized guidance information matching and publishing method,designs and develops rail transit passenger flow guidance APP based on theoretical results.It provides theoretical and technical reference for improving the level of rail transit passenger information service and operation management.The main research contents of this paper are as follows:(1)A travel preference analysis method based on passenger portraits is proposed.Aiming at the problem of passenger travel preference analysis and extraction,a travel preference analysis method based on passenger portraits is proposed.First,by analyzing AFC data,mining passenger travel information,and establishing passenger profiles from two aspects: objective travel characteristics and subjective individual characteristics.Second,the spectral clustering algorithm is used to classify passengers from the OD level,and an OD-based passenger preference identification method is proposed to extract and characterize different types of passenger preferences.Finally,the effectiveness of the method is verified by examples with multi-path no-transfer OD and multi-path multi-transfer OD.The results show that the OD-based passenger preference identification method can extract and analyze the preference information of passengers in different ODs and different travel periods.(2)A guidance information matching method based on passenger preference is proposed.Aiming at how to recommend guidance information to passengers according to their preferences,an guidance information matching method based on passenger preferences is proposed.First,the preference of passengers for different route attributes is considered in the route choice model,and a lexicographic preference model based on JND is proposed.Second,a model parameter optimization method based on reinforcement learning is proposed to achieve personalized guidance.After comparative analysis of examples,the recommendation accuracy of the lexicographic preference model based on JND is high,and it can better meet the personalized needs of passengers.(3)A multi-scene guidance information release method is proposed.First,the spatio-temporal publishing method of guidance information in crowded scenarios is determined.Second,in response to the passenger’s demand for guidance information in different scenarios such as emergency scenes and last train scenes,a guidance information release method based on multi-scenario is formed.(4)Based on the above theoretical methods,design and develop a rail transit precise guidance APP.The APP is divided into client and server.The client is an APP application that provides guidance services to passengers.The server includes the background services that support the APP application,the database,and the background system used by the operators.At present,APP application and background system have been put into use in Guangzhou Metro,which shows that the theoretical method proposed in this paper is feasible and applicable.