The Traffic Prediction Improvements (TPI) tool calculates the route an aircraft is most likely to fly, not by looking at the filed flight plan, but by looking at historical flight data.
One source of inaccuracy in flight plan based predictions is that the future clearances given by air traffic controllers are not reflected in the predictions: very often we ‘know’ that controllers will give permission to fly shorter routes, but until the input is actually made in the system, the prediction will reflect the filed flight plan. By augmenting traditional trajectory prediction logic with machine learning, a considerable improvement to accuracy can be achieved.
TPI is also set to provide predicted trajectory information to the Advanced Air Traffic Flow Capacity Management Planning Function (AAPF), a new position in the Maastricht sector groups under trial to detect potential traffic clusters up to 30 minutes in advance.