Multi-objective fuzzy rule-based prediction and uncertainty quantification of aircraft taxi time

The ever growing air traffic demand and highly connected air transportation networks put considerable pressure for the sector to optimise air traffic management (ATM) related performances and develop robust ATM systems. Recent efforts made in accurate aircraft taxi time prediction have shown significant advancement in generating more efficient taxi routes and schedules, benefiting other key airside operations, such as runway sequencing and gate assignment. However, little study has been devoted to quantification of uncertainty associated with taxiing aircraft. Routes and schedules generated based on deterministic and accurate taxi time prediction for an aircraft may not be resilient under uncertainties due to factors such as varying weather conditions, operational scenarios and pilot behaviours, impairing system-wide performance as taxi delays can propagate throughout the network. Therefore, the primary aim of this paper is to utilise multi-objective fuzzy rule-based systems to better quantify such uncertainties based on historic aircraft taxiing data. Preliminary results reveals that the proposed approach can capture uncertainty in a more informative way, and hence represents a promising tool to further develop robust taxi planning to reduce delays due to uncertain taxi times.