Aircraft Flight Status Prediction Methods

Along with the rapid development of China’s civil aviation industry, as well as the increasing traffic, airspace was restricted severely and contradiction between airlines and passengers was became more serious than before. These all have led to the controllers need to work under a greater workload compared to the past. Therefore, the use of the scientific method to predict the operation of the aircraft is essential to improve airspace efficiency, and protect the safety of aircraft operations. In this paper, aiming to solve different phases of flight status recognition and prediction problems, the model of flight mode identification in terminal area based on priori control experience was established. And we used learning algorithms to predict aircraft en-route flight phase of flight. Then prediction models in conflict situations are designed. Qt platform was used to develop digital flight strip software which contained flight mode prediction. It could provide effective theory approach and efficient aids to improve the efficiency of control. This thesis mainly includes:1. For flight mode identifying problem areas near terminal, we applied regular expressions to extract key information in track data. Then this thesis established a priori control experience PHMM model which was applied to identify the aircraft flight mode changes in approach and make a reasonable assessment to controllers’ work.2. For flight mode prediction problem in route phase, take advantage of the timing modeling capabilities of Hidden Markov Model and the predictive inference capabilities of feed forward neural network, a HMM-BP model was established. By learning to track a large number of historical data, the model could predict flight status changes.3. For the problem of two aircraft flying conflict situation, combining with experience and flight parameters, we improved the traditional Swarm model to predict changes in conflict situations on the basis of the potential function and Steer algorithm, then this model was simulated and validated by Matlab.4. We designed a digital flight strip software contained flight mode prediction. Based on the principles of software engineering, we design conceptual model using UML in line with national standards which specified detailing standards. Then to achieve a prediction function combined with strip’s operation, Qt platform was used to design the main interface and function of each part.