Detection And Release Of Flight Conflict Based On Artificial Intelligence

Flight conflict detection and relief is an important link to ensure the safety of air transportation.With the gradual opening of airspace and the increase of air traffic flow,more accurate,timely and efficient flight conflict detection and relief system will be needed.In order to solve this problem,support vector machine,case retrieval and multi-layer forward neural network are used to study flight conflict detection and flight conflict relief.Aiming at the problem of flight conflict detection,a model of flight conflict detection combining support vector machine and cylinder flight protection area is proposed.This model combines ADS-B data of two aircraft with flight trend module to predict the movement trend of the aircraft,and uses support vector machine to give accurate warning of flight conflict.The example analysis shows that the model has good flight conflict detection ability.A collision detection system combining improved cylinder protection area and multilayer neural network is proposed,the system uses improved cylinder model to divide the space around the aircraft into collision area and buffer area,constructs database to store and process ADS-B data,trains multilayer neural network with processed data,so that the neural network has the ability to predict the time when the flight conflict will occur At the same time,it can provide more accurate conflict time.By analyzing its confusion matrix,we can see that the system has good conflict time prediction ability.To solve the problem of flight conflict relief,a flight conflict relief system based on case retrieval is proposed,which collects and processes ADS-B data sets that have already had flight conflict,constructs a database with specific flight relief methods,and constructs a flight conflict relief case database.The new flight conflict problem is transformed into a target case.By calculating the matching degree between the target case and the source case in the flight conflict relief case base,the optimal case is retrieved,and the corresponding relief method is given.The target case istransformed into a new source case,then compared with the source case in the flight conflict relief case base,and finally the case base is updated.The example analysis shows that this method can provide a good solution to the flight conflict.When a new flight conflict occurs,the ADS-B data group is preprocessed and input into the multi-layer neural network,so as to provide a specific flight conflict relief strategy.The multi-layer neural network can give the command of changing the flight speed or the up and down speed to the flight conflict.