CISUC

The Role of Context in Transport Prediction

Authors

Abstract

After a few decades of research and development in intelligent transportation systems (ITS), we have an impressive amount of hardware and software tools that can monitor, estimate, and control the traffic network. Such tools are essential for traffic management and traveler’s decision making, but the complex role of human behavior in the transportation system demands considerations that might not be captured with sensors that are focused on the network or vehicles. For example, the traffic manager needs to understand why certain congestion is formed (Is it an incident? A special event? A religious ceremony? Weather? School pick-up/drop-off?) and to predict how it will evolve. A special event leads to different patterns and management procedures than an incident or a flooding event. In other words, besides knowing that a problem exists, traffic managers and prediction systems need to know its context.
So, what’s missing? How can we extend current ITS technologies to capture and process such information? Here, we suggest that the Internet is a resource for contextual information and we’ll overview available techniques and open questions to use it in ITS, particularly transport prediction.

Keywords

Intelligent Transportation Systems, machine learning, traffic prediction, context mining

Subject

Intelligent Transportation Systems, machine learning, traffic prediction, context mining

Journal

IEEE Intelligent Systems Magazine, Rosaldo Rossetti, February 2014

PDF File

DOI


Cited by

No citations found