B0589
Title: Interpolation of weather conditions in a flight corridor
Authors: Gong Chen - Dresden University of Technology (Germany) [presenting]
Hartmut Fricke - Technische Universität Dresden (Germany)
Ostap Okhrin - Technische Universitaet Dresden (Germany)
Judith Rosenow - Dresden University of Technology (Germany)
Abstract: Economic price pressure on airlines requires highly efficient air transport operations. On the other hand, current research initiatives, such as the Single European Sky Air Traffic Management Research Program request a future air traffic system with increased safety, efficiency, and environmental compatibility. Therewith, multi-criteria aircraft trajectory optimization with reliable meteorological information is becoming increasingly important in everyday operations. Data from the Global (Weather) Forecast System are provided at a resolution (28km, 6 hours) that requires interpolation to optimize trajectories with sufficient accuracy (about 200 m, 1 hour). For aerodynamic crucial weather variables such as temperature, wind speed, and wind direction, we investigate different interpolation models such as linear interpolation, Kriging, radial basis function, neural network, and decision tree regression with bagging and boosting. All methods are compared concerning cross-validation interpolation error and computation time. Considering an example trajectory from Prague to Tunis, Monte Carlo simulation is applied to examine how the errors in GFS data and the Kriging interpolation method can have an impact on the simulated trajectory. The results can be used for reliable, in-flight trajectory optimization, where small-scale changes in weather data become highly sensitive input variables.
