Use cases

Validating probabilistic network traffic forecast

Traffic demand prediction is key to enable all network users to act and react to traffic imbalances. To improve traffic prediction, a probabilistic approach has been developed where flights are applied probabilities to be still counted in the sectors or not in the coming hours/minutes. The assessment of this probabilistic traffic demand prediction required new monitoring indicators and tools to be defined and agreed by flow management positions (FMPs).

Trustful and meaningful prediction is key

In the context of the SESAR PJ09, the probabilistic demand prediction approach was defined by several FMPs and EUROCONTROL. It was pursued with the analysis of past Network Manager (NM) traffic and of the 65 busiest European airports in terms of movements, to generate uncertainties and translate them into probabilities on each flight to be in a sector at a given time.

“The development and assessment of performance of probabilistic demand prediction is an important milestone in the DCB domain, as it paves the way to more accurate traffic prediction used in the assessment and resolution of DCB imbalances.”

Peter Choroba Project leader, EUROCONTROL

New human-machine interface

The assessment of this probabilistic traffic demand prediction required new monitoring indicators and tools to be defined and agreed by flow managers. A new human-machine interface (HMI) was designed to represent the probabilistic values and enable their comparison with the current traffic measurements.

Using a realistic validation environment

The new HMI was assessed with simulations performed on the Network Manager Validation Platform (NMVP) running in shadow mode with live traffic. Post-simulations questionnaires and data analysis provided complementary results to this validation. Simulation sessions took place in Reims ACC with DSNA FMP and in the EUROCONTROL Experimental Centre with NATS and ENAIRE FMPs.

At a glance

4

simulation days

138

simulated flights

5

flow management positions

38

hotspots

Interested in our NMVP solution?

Consult our dedicated page, or contact us for any question.

Latest highlights