Regresar

Data-driven solution for planning bus routes of the public transport in UNICAMP

Abstract:

Managing the public transportation system today is a hard task and will continue to be more difficult as population rise. City centers are congested, air pumped full of noxious vehicles fumes, and public transport buckles under increased demand. Intelligent Transportation Systems (ITS) will attack these problems by relying on large databases, ubiquitous computing and communications to augment existing infrastructurebased deployments. The University of Campinas (UNICAMP) is currently developing two projects of great impact on the community, namely, the Sustainable Campus and Electric Mobility. Both projects have monitoring and control systems through various meters and sensors installed throughout the campus and bus fleet. That tsunami of collected data creates a suitable environment for data-driven applications to improve the internal public transportation system of UNICAMP. Machine Learning approaches are being investigated as one possible solution to transportation system issues. This work shows that classifying bus-stop based on demand can be very useful for improving bus routes and, consequently, the quality of bus service. In order to classify the bus stop demand, this paper uses the clustering with the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm applied to actual bus GPS data. From the bus-stop demand, more adequate routes can be proposed, even, dynamic routes can be adopted considering the demand changes along the day. An electric bus will serve these proposed routes in order to relieve the demand for transportation on campus. The proposed routes are analyzed considering the energy consumption of the electric bus and its autonomy.