Using wireless mesh networks to enhance journey time predictions in urban areas

Case Study

This project uses innovative embedded algorithms in wireless mesh networks combined with novel Machine Learning techniques to advance the current state of the art for real time journey time predictions in urban areas over a continuous rolling time horizon. This is a DfT sponsored project under a T-TRIG grant, working in partnership with IDT and Hertfordshire County Council.

The project explores the potential for using wireless mesh networks in conjunction with the Traak Analytics Platform (TAP) to provide accurate journey time predictions in urban areas over a 30-minute rolling horizon. The target accuracy of the system will be in the region of 90%. This level of accuracy over this time frame enables traffic managers to modify traffic network strategies to ensure that journey times are maintained at a consistent level by enabling the reduction of congestion and thus maintaining traffic flows across the network.

The system uses Machine Learning which ensures that the accuracy of predictions continuously improve over time. The data used is primarily from wireless mesh networks provided by IDT. This is augmented using traffic flow data from the UTC system of Hertfordshire County Council through their UTMC system.

Later, flow data can also be processed from ANPR cameras on the network. This can then be used to predict other factors such as Air Quality through a link to DVLA which can identify vehicle types and potential emissions levels. The system uses the latest AI and Machine Learning techniques which are further processed via the Traak algorithms. At a future date, journey times can also be published to other stakeholders using mobile apps and web services.

Interested in this product? Get in touch for more information