Flight Flow Prediction Based On ADS-B Data

In recent years,with the rapid development of China’s civil aviation industry,the number of domestic civil aircraft is also increasing rapidly,and the flight delay caused by air traffic congestion is becoming more and more prominent.Therefore,how to forecast the air traffic flow has become an urgent problem to be solved.ADS-B technology has become a new way to study air traffic flow prediction by virtue of its characteristics of large amount of data and rich data content.It is of great significance to conduct in-depth research on historical ADS-B data.To forecast the flight flow is the key of flight track forecast,this paper studied the prediction method based on circle/rhumb line track,based on intention to flight track prediction method,based on grey model to predict the track of law of the three traditional track prediction algorithm,numerical example experiment was carried out respectively,and the experiment result error analysis.Secondly,an improved spatio-temporal track clustering algorithm was proposed through theoretical research.K-means++ algorithm was used to cluster the time data,and similarity was calculated according to the modified Euclide-distance algorithm.Sampling method and fuzzy clustering method were used to cluster the track in the straight stage,and interpolation method and DBSCAN clustering method were used to cluster the track in the turn stage.The prediction track can be obtained by modifying the current position of the clustering track,and the error result is analyzed by the example test with the actual track data.The results show that the algorithm has a good prediction accuracy.Finally,the proposed improved spatio-temporal flight path clustering algorithm is applied to air traffic flow prediction,and a flight flow prediction system within the scope of Mianyang Airport is designed and implemented.Through the experiment,the incoming flow,departure flow and overflight flow of Mianyang Airport are predicted,and the predicted results are compared with the actual flow values.The results show that the system has a higher prediction accuracy,and can provide data basis for the controllers to manage the flight flow.