In this paper, we consider a problem of management of crisis situations (incidents on nuclear or chemical plants, natural disasters...) that require remote sensing, using a set of ground and aerial robots. In this problem, sensed data must be transmitted in real-time to an operation center even in case of unavailability of traditional communication infrastructures. This implies that an ad hoc wireless communication network must be deployed, for instance using a fleet of UAVs acting as communication relays. From a technical point of view, we tackle a scheduling problem in which activities of mobile sensing robots and mobile relays must be synchronized both in time and space. Schedules produced must also be flexible and robust to the uncertainty about the duration of robot moves at execution time. The problem is modeled and solved using constraint-based local search, with some calls to graph algorithms that help defining good communication networks.
展开▼