Research Of Bus Travel Time Calculation And Prediction System Based On GPS Data

City bus transportation is an important part of traffic. In order to meet traffic growth in travel demand and enhance the social benefits of the bus traffic, it is necessary to further improve the effectiveness of bus traffic and require a more efficient scheduling of operators of vehicles, to provide a more accurate vehicle arrival information, which is its core Bus travel time prediction based on a higher demand, the need for more accurate and rapid travel time prediction method and system to support.Bus Travel Time Prediction on the methods of research, reviewed an overview of relevant research at home and abroad. By analyzing the characteristics of public transport as well as the formation time of the impact of factors buses and bus journey times based on factors (such as road travel time, delays intersection, bus stop in time) and travel time prediction system for the data needs (static information, quasi-dynamic information and dynamic information) summed up the model of the whole project basis, as well as the establishment of specific methods. Base on the above analysis of historical data based on the return of Dynamic Travel Time Prediction Model. The model include: data pre-processing module, this module from the time check, check road track, the speed of the three aspects of calibration data to determine whether the normal GPS can be used; normal prediction module, according to GPS data contained in the current vehicle location and feature points on the lines of latitude and longitude coordinates to calculate the current location of vehicles to the various e-stop distance, and then based on historical average speed data to calculate the various e-stop journey time forecasts; abnormal prediction module to estimate the current bus position, calculated in accordance with its coordinates to all E-stop distance. The model is able to return to the normal data and abnormal data processing logic must be able to predict the arrival time. Bus operators through the process and the impact of travel time factors analysis to Kalman filter recurrence formula as the basic method, the use of advanced technology transit bus automatically obtain location information and the number of passengers up and down information, to predict bus travel time , in stops time, the bus journey time to establish an integrated model projections, especially for single bus in the intersection signal control of the delay time for the analysis of the basic model be amended accordingly.Further delineation of the system and functional module; on the data receiving module, the initial database module, database storage and processing module, bus arrival prediction of functional modules; given the composition of the various modules and functions of the class description, describes the various types of methods and interfaces; introduced a system of database design and interface design. During the design process, using the method of object-oriented design patterns in combination with the design of the system in order to reduce the coupling system to improve the system scalability. Extraction system with universal module design, allows the realization of their business as much as possible of the loose coupling, thereby enhancing the module's versatility and reusability and can be customized to meet user requirements.In the implement of system, the continuous optimization of the system to improve the demand for change in the design of exposed deficiencies. Not only to optimize the system architecture and details of the technology to optimize and improve the system's maintainability and efficiency. System of local data receiving module will send the data center model for processing. If the transmission of data from static information will Xml string corresponding value of all the secondary node removed, saved to the array, and then the corresponding three-node out of all the values, save it to another array. If the transmission of data from real-time data, will be designated under the two nodes at all three levels of the value of the node removed, saved to the array. The initial database module for the bus to monitor the database in different static data structure, the establishment of the corresponding initialization SQL scripting language, as input to initialize module. Module - based model of bus arrival prediction algorithm needs data structure, the establishment of the standard database (Standard DB), the establishment of standard library initialization script as the input module. Bus stop out front of the forecast module has been deposited in the database of legitimate data, including static information and dynamic information, and then remove the data to deal with, first of all to determine whether the data meet the processing conditions, because there is no need for some data processing, can retain the original value (the procedure can effectively reduce the computational complexity), if the data processing to meet the conditions of the process according to the procedures for processing of data.Finally, 49 bus lines in Shanghai measured data, this paper, the bus travel time prediction model is validated analysis and evaluation, and summarizes the characteristics of the model, real data and forecasts through data comparisons between to verify the model. Model shows the correctness and the implementation of the system can be.The system has been delivered to users, and in accordance with a phased implementation strategy into practical operation, the user response is good, that the system meets the user's actual requirements.