Relating different types of (traffic) data

Creating a congestion search engine

This process allows a user to use search queries based on congestion patterns, and find all relevant similar situations where a chosen traffic pattern occurred. This creates the possibility to ‘reverse-search’ the database, instead of looking at specific congestion patterns on a selected time, date and route. This potential increases the usability of NDW-data and makes innovative analyses possible.

This CoZi (Congestie-Zoek-Engine) will be developed in the first half of 2016, making it possible to present results this late spring).

This completes the first research project undertaken by DiTTlab for NDW, which sets the standard for new data storage, relation and analysis possibilities for the users of NDW data.

Multi-scale traffic state estimation

Potential

Starting in 2016, DiTTlab will be elaborating the next phase of the NDW-project. Using the NDW-data and combining this with many other potential data sources, the goal is to allow:

• Estimation of unknown traffic parameters, like traffic volumes, vehicle-loss hours, inflows, turnrates, capacities, critical speeds and fundamental network diagrams;
• Estimation of origin-destination matrices on different spatial levels, from road lane to the macroscopic regional level;
• Simulation and prediction of traffic on the basis of all available data for the different spatial levels.

This will create the potential for new applications for traffic simulation, prediction and a large increase in data relations leading to novel insights and analyses.

Multi-scale spatial levels

Each spatial level requires a specific and relevant input for an optimal analysis. Clearly, the required input differs for each spatial level. The table below illustrates the various spatial levels and their corresponding data input.

The multi-scale approach will enable an accurate and reliable estimations of the prevailing traffic state that will help efficiency and effectivity of all applications, from personal traffic apps to traffic control centers. This project will resolute in a framework that will be based on a fusion engine to incorporate all available data sources (based on the aforementioned initial NDWproject), and turn these into one coherent view of the traffic state. DiTTlab plans to work from an open source perspective, to create a framework available for all to build upon further. This approach makes the project unique in the world, allowing the results to be used in many existing and new applications within the ITS community an transportation in general.