Study On Real-time Prediction Of Arrival Time For Floating Transit Vehicle

Bus arrival time prediction is one of key technology to realize intelligent public transport system. Accurate and reliable bus arrival time prediction is not only to enhance the level of public transport services, but also to supply managers with the basis of scheduling and improve efficiency of bus operation. This thesis aims at forecasting bus average travel time, bus stations and signal lamps region delay time, and then establishing the real-time prediction model of bus reaching the downstream stations. Through verifying with real-time GPS data of Chongqing City bus and providing basis for the process of urban public transport information building. The works of this thesis are reflected as follows:In order to calculate the road sections of bus average travel time, it is that judges the effectiveness of floating vehicle data by fuzzy comprehensive evaluation method. At the road section with sufficient data, it calculates the average travel time by instantaneous velocity replacement and compound integration methods; while at the road section with insufficient data, it considers the adjacent sections GPS data of objective section, and improves the calculation credibility of average travel time by endowing the corresponding weight in accordance with its contribution.And then, through analyzing the whole process of bus stop, discussing the factors of bus deceleration and the method of judging the time of bus deceleration. Taking into account the effective of bus stop location and time, the delay model of boarding and alighting is modified. Moreover, it illustrates that bus acceleration of out station from straight and harbor bus stations. On this basis, the bus delay model at the station is established.According to analyze the factors of bus delay time at the signal region, it describes the random process of bus passing the signal lamps area by using typical distribution of mathematical statistics, and proposes the assumption conditions of bus passing by the signal lamps area, then the bus delay model at signal region is established by introducing queuing theory.On the basis of considering the partial operation condition of urban bus,it is proposed that the bus arrival time to the downstream station model in real time. The model was tested and verified with real-time bus GPS data in Chongqing City, the results indicate that the bus arrival time to the downstream station can be predicted in real time, the accuracy is better than existing methods, the model has better adaptability for various of traffic conditions. It can provide the higher precision of prediction, and the prediction errors fluctuation maintains a low level.