Research On Four-dimensional Track Prediction Technology Based On Data Mining And Deep Learning

With the development of China’s air transport industry,the rapid growth of flight flow and the sharp increase of all kinds of data in the air traffic control system,it is an inevitable trend to mine the organization pattern and evolution law of traffic flow from complex data and apply it to the field of flight track analysis.Accurate track prediction technology is becoming more and more important to solve the problem of airspace resource shortage,and it is also one of the key technologies of air traffic control based on track operation in the future.Aiming at the problem of precise four-dimensional track prediction in terminal area,a four-dimensional track prediction method based on data mining and deep learning was proposed,and a data mining module and a deep learning module were constructed.The data mining module can mine and automatically identify the approach mode of aircraft.Based on the results of pattern recognition,the deep learning module realizes dynamic precise four-dimensional track prediction.Firstly,the relationship between the multi-source data generated by the aircraft operation is analyzed,including the track data of the secondary radar system,the flight plan data of the CDM system and the meteorological data in the message.Then,a track clustering model based on D-CURE algorithm is constructed,and 19 approach patterns are mined from nearly 3 million track points.Then,the random forest algorithm is used to predict the track classification,and the pattern recognition accuracy is up to 96.36%,which changes the current situation of low accuracy of traditional algorithms.Finally,aiming at the complex problem of deep learning parameter tuning,whale algorithm is used to select the parameters of LSTM,establish the track prediction model,and train the data sets under different approach modes respectively.The example shows that RF-WLSTM algorithm has the best prediction effect and can achieve precise track dynamic correction prediction.It can provide basis for air traffic control and flight flow management.Figure42;Table14;Reference 51...