Today's smartphones contain a number of sensors (e.g., GPS, accelerometer, camera) that can be used to reveal driver activity and traffic patterns when carried in a vehicle. This project aims to gather a large-scale set of smartphone sensor data from volunteers while they are driving. The data set and the tools we will provide for analysis and access will have an impact on multiple areas of research, including data mining, vehicular networks, intelligent transportation systems, and traffic engineering. Our first step in gathering such valuable data is to determine the particular types of data, tools and scenarios that would be most valuable to the research community.
We plan to develop the DriveSense platform, a collection of tools for large-scale traffic and driving activity data collection. This will be accomplished through a large-scale deployment of a free smartphone application to collect data from a wide range of individuals, locations, and driving situations. We divide the components of DriveSense into platform components (software, tools) and management components (recruitment, data management). This will provide a rich and complex data set for use by the research community. This will support transportation research on the fine-grained information about traffic and driving behaviors collected by various sensors of a driver's smartphone, as well as enabling advances in more general research into the analysis of complex, spatio-temporal data.
The first step is to explore, develop, and validate specific components of the DriveSense platform in order to reduce risk involved before beginning full-scale development. More specifically, during this planning phase of the project, we will: i) Develop and validate several components of the front-end application that will run on smartphones to collect data. We will develop a prototype of the DriveSense App that collects data for the Android platform. ii) Hold a workshop to get feedback on research community interests and needs. The objective of the workshop is to bring together experts from several research communities spanning computer and communication networks, transportation systems, and data analytics to discuss community needs and interests, challenges to be addressed, tools to be developed, and new research problems utilizing the driving activity data.
The outcome of the planning phase will ensure we are able to capture the types of data most important to the community, so that the developed infrastructure will be of most benefit to the broader research community.
