Dynamic Aircraft Conflict Alert Based on the Markov Decision Process and Simulation of Actual Flight Data

Abstract

As air traffic flow continues growing, the risk of collision between aircraft is also increasing as well. In order to ensure efficiency of airspace would not reduce due to the increase of the conflict, and thus the accuracy of alerts also has higher requirements. Traditional conflict alerts based on the static threshold method will have a great flight separation redundancy, resulting in reduced operating efficiency of airspace. So it’s necessary to study the dynamic threshold alerts approach based on actual flight scenes, which can dynamically adjust the threshold and improve operational efficiency. First in this article, Markov decision processes were introduced. And then proposed a dynamic threshold alert method based on the Markov decision process. Finally, the analysis of alert results were performed through simulation of actual flight data. The results show that this method can reduce the false alert rate and improve the performance of the alert system under the premise of ensuring flight safety.