A Statistical Approach for Taxi Time Estimation at London Heathrow Airport

With a predicted year on year increase in the number of flights [5], increasing attention will need to be paid to the environmental effects of air transportation, in addition to an increasing focus upon maintaining and improving on-time performance at airports [3]. These challenges will face both airlines and airports in the near future and improved taxi time predictions (which is the aim of the model discussed here) will be an important tool for handling them. The purpose of the ground movement problem is to guide aircraft around the surface of the airport to arrive at their destinations in a timely (and potentially more environmentally friendly) manner. From an optimisation point of view, the ground movement problem can be considered to be one of the most important airside operations at an airport [2], since it links together several other important problems, such as runway sequencing (for arrivals and/or departures) and gate assignment. The aims of this combined routing and scheduling problem are usually to reduce the overall taxi time, reduce the fuel burn, arrive at the destination by a target time, and/or to absorb any necessary waiting time or delays in preferred positions (e.g. before the engines are started). It has to guarantee conflict-free routes for each aircraft throughout the movement and, therefore, usually has to coordinate the movement of multiple aircraft around the surface at once. A comprehensive review of the problem can be found in [2]. In previous research, it has been difficult to quantify the benefits of new ground movement approaches compared with the status quo at airports, since it can be hard to distinguish between the effects of the variability in taxi times and the benefits of the improved routing and sequencing. Research is badly needed to understand and quantify the variability in taxi times, and moreover, to develop models which accurately predict taxi times from more measurable but influential factors. In particular, if predicted taxi times are to be used in a decision support tool for the ground movement problem, which itself incorporates the effects of re-routing and queuing delays, the predicted taxi times must not already include the effects of these. The production of a model to quantify and eliminate these delays has not, however, been simple. London Heathrow Airport is one of the busiest international airports in the world, despite the fact that it operates with only two runways and (for noise control reasons) is restricted to using only a single runway at a time for departures. There has been