Conflict Risk Assesment Of Air Traffic Operation Based On Multi-agent Simulation

In recent years, with the increasing air traffic demand and limited air traffic control capabilities, the contradiction between demand and supply is looming large. In such context, it is necessary to conduct risk assessment whether adjust the existing structure of airspace or introduce new technologies and new concepts. As air traffic system is a complex system, computer is an important research tool to simulate the air traffic operation. However, existing simulation tools have two drawbacks, one is just one computer is not competent for large-scale microscopic air traffic simulation, the other is lack of mutual communication between air traffic entities. For these two problems, we applied Distributed Artificial Intelligence- multi-agent technology in air traffic simulation and developed the multi-agent based air traffic simulation system to identify and evaluate the conflict scenarios.Based on the research of the theory and application of agent technology, this paper first constructed the basic framework of multi-agent based air traffic simulation system, including the detailed design of aircraft agent, ATC automation system agent and air traffic controller agent. Secondly, a trajectory generation model was proposed based on BADA and aircraft intent. Through establishing a standard climb, decent and cruise profile, a detailed description of the aircrafts’ horizontal movement, vertical movement, the change of thrust, drag, configuration, fuel consumption was made from takeoff to touch down taking the weather into consideration. Thirdly, based on the study of the air traffic controllers’ behavior, a conflict detection algorithm was proposed based on the flight intent. The conflicts were classified according to ICAO regulation, several conflict resolution strategies were designed for each conflict category. Moreover, a controller intent recognition algorithm was put forward based on history-data clustering. The result can be used to design alternative route for aircraft spacing and sequencing in terminal airspace. Then, based on the system design and modeling of the key technologies, a prototype system was developed using JADE platform and Java language. A distributed structure was applied in the system which can simulate the whole day operation of the flights in China. The conflict data can be recorded during the operation of the system. Finally, we analyzed the conflict distribution in macro view from the aspects of category, time and sector. The potential risk scenario is identified through analyzing the recorded conflict data. For the specific risk scenario, a micro conflict risk assessment model was put forward to capture the conflict evolution process, and four measure index including conflict rate, conflict severity, conflict risk, total conflict risk were proposed.