The Design And Implementation Of Bus Arrival Time Prediction Methord In Real_Time Bus Information Service System

With the rapid development of urbanization, the population of cities is growing and the scope of life is expanding. The problem of people’s lives emerges under the travel pressure. Both in developed and developing countries, ITS(Intelligent Transportation System) is one of the indicators of economic growth, and APTS (Advanced Public Transportation System) is an important part of ITS. The rapid construction of APTS can largely alleviate the current traffic from the existing range of issues. Bus travel time prediction is a very important part of APTS and the high prediction accuracy can improve the overral level of ITS.To address these problems, the authors divided the work into two parts, which is real-time positioning and run-time prediction. In real-time positioning, the author proposed a method using base stations location and A_GPS location, it could get the real-time bus location information. In run-time prediction, by studying the current existing prediction models, a hybrid model based on Kalman filter and Elman neural network is proposed. This hybrid model takes both the historical data and the real-time data into account, so that it can predict the link travel time on the next time interval accurately. In the experiment, the author validates the proposed model and analyzes the error using real data. The results show the high precision of the proposed model.After the detailed technical solution of the key issues identified, the two approaches are applied to real-time bus information service system. The overall detailed designation is completed through the requirements analysis. The function of bus arrival time prediction is divided into three parts due to that the historical data is not fully true and the collection is very limited. The simulation data was used as the data base of the link travel time prediction during the first two stages. The real data was used for prediction in the third stage after the system was put into use. Finally the test results were shown and the recommendations were made after the analysis of the system deficiencies. Meanwhile the author summarizes the work in the period as a graduate student.