Single-Engine Taxi Intelligence: Questions and Answers
How TripAI helps airlines identify practical SETO and SETI opportunities, update them as airport conditions change and measure the result after flight.
What are SETO and SETI?
Single-Engine Taxi-Out (SETO) means an eligible aircraft taxis with one engine operating for part of the journey to the runway. The second engine is started early enough to satisfy the airline's approved start and warm-up requirements before takeoff. Single-Engine Taxi-In (SETI) means an eligible arrival may use one engine for part of the journey to the stand, after the applicable cooldown requirement has been met. Exact procedures vary by airline, aircraft, engine and operating condition, and not every flight is eligible.
Airlines already know single-engine taxi. What does TripAI add?
Airlines already hold the procedures, the approved requirements and the operating authority. TripAI makes the opportunity flight-specific. It predicts a practical window for one aircraft on one taxi, updates that window as the airport operation changes, explains why an advisory is shown, revised or withheld, and creates a post-flight record of what was available and what happened. It does not replace airline procedures, dispatch authority or crew judgement.
How do aircraft physics and machine learning work together?
Approved aircraft and airline requirements define the operational boundaries. Route geometry, gate, expected runway and airport context establish the physical journey. Machine learning (ML) then predicts taxi duration, remaining time and congestion effects, which are the parts of the problem that genuinely vary with conditions. Predictive models do not modify warm-up, cooldown or eligibility rules, and no advisory is issued when confidence or data quality is insufficient.
How does TripAI predict congestion?
Not all additional taxi time is congestion. TripAI first establishes a route-specific operating baseline from the gate, expected runway, likely route and plausible distance. Predictive models then consider surface traffic, departure and arrival queues, runway configuration, weather, recent movement and comparable historical operations to estimate what the current situation may add, and the remaining taxi time is updated continuously. Calibration is airport-specific, and every prediction carries uncertainty.
Is the advisory fixed before taxi begins?
No. It begins as a planning estimate built from the information available before movement. Aircraft progress and airport conditions then change: queues build or clear, routes and runways are reassigned, and the aircraft may stop or move faster than expected. The window can be updated, shortened or withdrawn accordingly. Stale data or insufficient remaining time produces no advisory at all, and the output remains advisory throughout.
Does TripAI tell the crew when to start or shut down an engine?
No. TripAI provides an airline-configured advisory opportunity. It does not operate aircraft systems or replace the airline's Standard Operating Procedures, dispatch authority or crew judgement. Airline procedures remain controlling at every point, the crew retains authority over the aircraft, and operational deployment requires airline approval and validation first. The system describes an opportunity and its reasoning, and people decide what to do with it.
Can ADS-B or surface movement data prove that single-engine taxi was used?
No. Automatic Dependent Surveillance-Broadcast (ADS-B) and similar surface movement data can show the route, timestamps, speed, stops and observed taxi duration, which is enough to reconstruct where a modeled opportunity existed. It cannot show engine state, the exact engine start or shutdown time, actual fuel flow or a realised saving. Confirming that an opportunity was captured requires authorised airline engine and fuel records.
How is the fuel opportunity estimated?
From the eligible window and the difference in fuel consumption between the applicable operating modes, calibrated to the specific aircraft and engine using airline data. There is no universal saving rate. The difference varies by aircraft type, engine, weight and operating conditions, so a value taken from one fleet does not transfer to another. Until authorised engine and fuel evidence is available, the result is reported as a modeled estimate rather than a validated one.
What data does TripAI need?
It depends on the stage. Prediction can use flight identity, aircraft type and configuration, gate or stand, runway and route context, surface movement, weather and airport status, and the airline's eligibility and buffer rules. Validation is a separate question and needs engine-state and fuel data, which may come from Aircraft Communications Addressing and Reporting System messages, Out-Off-On-In milestones or Quick Access Recorder and Flight Data Monitoring records where authorised. Not every feed is required for every stage.
How can an airline begin safely?
Begin with historical replay for one aircraft family at one airport or operating cohort, which establishes where opportunity genuinely exists. Move next to live shadow mode: TripAI evaluates and records the recommendation without changing the live operation or displaying an operational advisory to the crew. Predictions are then compared with airline evidence and reviewed by flight operations, engineering, safety and fuel teams. Expansion follows only after acceptance criteria are met.