Study on Flight Landing Time Prediction in Complex Terminal Areas

The impact of control deployment, and complex airport terminal area flight landing time uncertainty is greater. One significant way to ease this situation is to predict the flight landing time accurately. However, in complex airport regions, flights are more susceptible to various factors, such as control factors and airspace factors. In this context, effectively predicting flight landing times becomes more difficult. Aiming at this problem, this paper proposes to comprehensively consider various factors affecting the flight landing time in the complex terminal regions and use XGBoost regression algorithm to predict the flight landing time. Firstly, the variables affecting the landing time of flights in complex terminal areas are explained to determine the variables used in the experiment. Then input the data into the XGBoost regression model for training, take the remaining flight time of the flight as the target value, establish the flight arrival time prediction model, quickly realize the model training, and predict the expected arrival time of the flight in three different peak periods. It is found that the model has good control over fitting ability and high prediction accuracy; Finally, by comparing the prediction results using multiple variables with those using only longitude, latitude and height, it is verified that the prediction accuracy using multiple variables is higher.