Air traffic
control
research

Taxi-out as a record

At ten European airports the taxi-out time is the gap between two timestamps, and the off-block one is whatever each airport's system recorded. Pick one of five real departures and watch where the stamp lands: on the pushback, on the schedule, a day off, or before the aircraft landed.

MethodReal 2025 departures at ten airports; the Amsterdam routes are the receivers' own ADS-B tracks. Times UTC. The final model carries no value fitted to leaderboard scores.

Open experiment→

Ninety Seconds

London lands one runway from four holding stacks. However irregularly aircraft leave the stacks, they touch down about ninety seconds apart, over and over. The release order does not carry it, the wake step is muted, and the gap is rebuilt on final.

MethodThe delivered gap read from public aircraft motion, measured on short final over four westerly days. The release interval scatters several times as wide as the gap; a generic wake rule would make several bands. The delivery is one.

Open experiment→

Runway in use

Mexico City and Los Angeles: which runway each airport is landing and departing on, right now, read from the motion of the aircraft themselves. The estimate carries its own confidence and goes dark when it cannot see.

MethodLive configuration from public aircraft motion. Arrival or departure from vertical trend, the runway from lateral offset to each centerline. Confidence shown; offline when the signal is stale.

Open experiment→

Negative airspace

New York and Mexico City: as congested as airspace gets, but surely there's still room to maneuver? An active volcano marks the one hole that is real.

MethodObserved traffic on published arrival routes, counted within vectoring reach over a few days of traffic. Terrain and minimum vectoring altitudes not subtracted, so the room shown is an upper bound. Public inputs only.

Open experiment→

Controllable Airspace

Mexico City and New York: comparing sector geometry, control complexity, and airspace structure in two high-density corridors.

MethodEqual 40 NM disk, 11,000 ft ceiling above each field; terrain subtracted from the bottom. Public inputs only.

Open experiment→