A demonstration of dedicated constraint-based planning within agent-based architectures for autonomous aircraft

Autonomous agents are a challenging concept for future unmanned air operations in hostile environments. Previous aeronautic missions highlight the lack of on-board reasoning abilities to increase the decision making capability, to efficiently react to unexpected events or to adapt plans to unexpected situation changes. If many agent-based system approaches exhibit reasoning functionalities, the complexity of air missions prevents from using the underlying generic models onto realistic missions. By taking advantage of constraint programming techniques, this paper demonstrates how a dedicated planning method can manage unmanned air vehicles into a realistic mission.