A multi-agent approach to reliable air traffic control

International audienceAir Traffic Control (ATC) is going to be a typical critical socio-technical system in which controllers use a large number of distributed software tools to provide safety ATC services. The reliability of these services relies on the availability of the various tools. In the process of integrating more and more sophisticated tools in their daily work, controllers need to feel confident in the reliability of their tool set. This paper presents a multiagent approach to this reliability problem. We propose an agent-based decision-aided system that helps controllers in using their multiple software tools in situations where some tools are not available due to technical incidents. As it is critical to conduct experiments on real air traffic, we build and test our Multi-Agent System (MAS) in a simulation environment, thus develop an Agent-Based Simulation (ABS). Experimental work on this ABS has demonstrated the significance of our system to air traffic controllers