A Prediction Model for Bus Arrival Time and Multi-Route Travel Time

以不同的旅行者需求发展出适宜的公交车到站时间预估模式，并依预测时距分为长期与短期复合路线旅行时间预估模式。在长期公交车旅行时间预估模式的构建中，应用动态随机性旅行时间模式计算旅行时间的期望值与变异数，再利用快速傅里叶变换有效推估出函数型态，用于求取旅行时间的动态变化；而在短期公交车旅行时间预估模式中，则以前后站的车辆总延滞表现。最后，以实例依预测时距对预估准确度作图，找出长短期预估模式适用的分界点为：预估35min以内的公交车到站时间适用短期模式，而超过35min则采用长期模式。最后验证其在复合路线上的预估准确度。To fulfill the different demand of passengers, this study divides the prediction model for bus arrival time into a long-term sub-model and a short-term sub-model. In building the long-term sub-model, this paper applies dynamic and stochastic mode of travel time to calculate the expectation and variance of link travel time, and uses Fast Fourier Transform (FFT) to obtain a reliable function type. The delay levels of both passed and next stops one considered to obtain a robust short-term prediction model. This study also reveals that the applicable point of demarcation between the long-term and the short-term models in Taipei is 35 minutes. A trip type with transferring among multi-mutes is validated via a real case study