Chapter 10 - Bus Arrival Prediction and Trip Planning for Better User Experience and Services

Abstract

Public transport is regarded as one of the most efficient means to confront the explosively growing city traffic problems such as congestion and pollution. Generally, public transport is hard to be more adopted by citizens, because of its poor service. In order to provide good bus services and improve user experiences, in this chapter, we will focus on the technologies for bus arrival prediction and bus trip planning, which are the two key components for public transport system. First, a statistical approach is presented to predict the public bus arrival time. It describes the bus arrival time as a linear model while considering a number of factors such as departure time, work day, current bus location, number of links, number of intersections, passenger demand at each stop, traffic status of the urban network, and so on. The parameters of the model are trained by the historical traffic information. Second, a system of bus trip planning service is proposed, which can help public transport users choose the most appropriate bus lines and transfers based on real time and predicted bus arrival time. Because of the large scale of the data size, a two-step combination of K -transfer and Multiobjective algorithms is used to optimize user-traveling route. Finally, a prototype is built to verify the practicability and efficiency of the proposed approaches. Experimental results on real transport networks of large cities proved that the system is efficient and practical.