years. The small size and the relative ease of deployment of Fixed-wing MAV make them a very interesting support for military or civilian field operations. During the last five years fixed-wing MAV system have already demonstrated their capabilities of autonomous navigation. One of our current concern is autonomous decision making like swarm coordination or mission planning. Solving the mission planning problem for a swarm of MAV consist in allocating tasks to each to achieve all the goals of the mission and to schedule the execution of thoses tasks in the flight plan of each MAV of the swarm.
We propose a formal model of the online mission planning problem written as a Constraints Satisfaction Problem (CSP) that we solve it with the FaCiLe constraints library. However plans need to be recomputed online to take into account the uncertainties during the performance of the mission. Hence we provide a classical reactive architecture which ensure that the swarm will finish the mission in a finite time. This architecture has been implemented in the Paparazzi System.