Research Of Urban Rail Transit Passenger Flow Characteristics Based On Phone Signaling Data

Currently, the urbanization process of China is still in the period of rapid evolution, along with the expanding of urban land use and the rising of the population. The temporal and spatial travel scope of urban residents was expanded further and larger, however, the taffic problems was more and more serious. The urban residents benefited greatly from the urban rail, which have the advantage of large capacity, long-distance and so on. However, with the continuous expansion of urban rail transit network, difficulties for the operating department, such as large passenger flow warning, subway network layout optimization and the improvement of the operating organization, have emerged. To solve the difficulties, the subway passenger travel characteristics need to be acquiried timely and accurately. The analysis of subway passengers’travel characteristics relied on accurate and integral data, and the limitations of traditional data acquisition methods were obvious. As a new research direction, the sample size and randomness of the phone signaling data was ensured by the increasing number of mobile phone users and the increasing usage of mobile phone. On the other hand, the characteristics of the phone signaling data, such as time-continuity and high space coverage, was formed and reinforced because of the continuous expansion of wireless network coverage. Using the temporal and spatial information of the phone signaling data, the rail passenger travel characteristics could be analyzed and judged scientifically with quantitative method, and the real-time dynamic traffic information could also be digged out, so as to provide scientific and quantitative decision support for those difficulties.Based on the knowledge of the wireless communications network, the Cell of Origin positioning technology and the Handover positioning technology were introduced, in order to theoretically explain the generation process of the phone signaling data, as well as the temporal and spatial information contained. According to the record rules of the phone signaling data, the definition of each field in the phone signaling data was explained in detail. SQL Server database software was used to store and manage data, which contained the phone signaling data preprocessed and the data about the underground and ground station information of the subway GSM system and so on, for the purpose of establishing the data basis for the further analysis.Whether analyzing the subway passenger travel characteristics or mining real-time dynamic traffic information, the first critical step was to identify several important travel behaviors of the passengers and the associated travel path, using the temporal and spatial information of the phone signaling data. Based on the in-depth analysis of the features of the subway wireless communication network, the recognition principle of subway passenger travel path was propsed, and then an efficient process for recognizing the travel path was established, which contained preprocessing data, selecting subway travel data, establishing an algorithm for identifying the travel stations and judging the effectiveness of the travel path identified. Above all, the last two steps was the core part of the process for recognizing the travel path. The verified results, deduced by real data, show that the travel path of individual cell phone users in the subway could be recognized accurately and efficiently, with the help of the process for recognizing the travel path.On the basis of the process for recognizing the travel path, the characteristics of enter-station behavior, exit-station behavior and transfer behavior in the designated subway stations was analyzed, using one-day data on weekdays. First of all, the ideas of recognition algorithms for the three behaviors was put forward, that is, when suspected travel behavior characteristic data was filtered out, checked the LAC information of the previous line data whether it meets the requirements or not, if it meets the requirements, record the phone user, if not, abandon the user. According to this idea, a java program was programmed, and the time-varying maps of passengers flow in the designated subway stations were achieved. Then considering the land use situation around the subway station, the characteristics of enter-station behavior, exit-station behavior and transfer behavior in the designated subway stations was analyzed respectively. Research results show that, first, the peak phenomenon of enter-station passenger flow at different station is related to the land use situation around the subway station, second, the subway stations located in the commercial office district have obvious tide phenomenon and a large amount of activities at noon, third, transfer passenger flow among different lines shows tide phenomenon, and transfer passenger flow fluctuated little during the rest of the time period. Therefore, the subway passenger travel characteristics, matching with the land use situation around the subway station, can be obtained by using phone signaling data accurately.Considering the index of metro occupancy rate (MOR), fluctuation of MOR for the specific trips during peak period was evaluated quantitatively. Fusing the information of phone signaling data with subway schedule information, multiple sets of random enter-station time and exit-station time was achieved. Then, a time allocation model for the designated station was established with the tatistical analysis of these time data. On the basis of the time allocation model, the number of designated-station passengers, getting on or off the train, could be converted to enter-station passengers flow or exit-station passengers flow in a specific time interval correspondingly. Counting the enter-station or exit-station passengers flow in the specific time interval at each station during the designated trip, the number of get-on or get-off passengers at each station could be acquired by this method, so as to evaluating the fluctuation of MOR for a specific trip quantitatively by adding and subtracting the number orderly. The results which verified by the phone signaling data randomly selected during peak period, show that the accuracy of this method can up to 90%, which means that the method is capable to evaluate the MOR.