Research on the Dynamic Data-driven Application System Architecture for Flight Delay Prediction

Flight delay prediction remains an important research topic due to the dynamics of the flight operating process. To solve this problem, a dynamic data-driven approach from the control area has been introduced, in which real-time data was collected and injected into the prediction process to get more accurate and reliable results. In the case of predicting the landing delays of consecutive arrival flights, delay propagation was analyzed to establish the corresponding state space model. Then, dynamic datadriven prediction architecture for flight delay and the prediction steps of on this architecture were presented. Several experiments were carried out on historic flight data to validate the performance of this solution. Results show that the accuracy is high, and not sensitive to the number of the predicted flights. Therefore, the solution has good predictive stability and reliability.