Limited resources in ambient systems for disaster scenarios

We describe a concept of autonomous intelligent and ambient networks suited for disaster scenarios (e.g. pollution leakages, earthquakes or forest fires) where the decisions are made only on the basis of sensed data, without human control. We discuss possible system architectures: its passive component (sensors) that monitors the environment, as well as the active one, i.e. robots or other machines (actuators) that are able to react accordingly to the previously gathered data. We highlight some challenges regarding distributed and self-organizational aspects of these networks. Also, we address an issue of collaboration that should exist between sensors and actuators. The decisions about the actuator operations are made based on previously sensed data, but these operations can also improve the quality of information gathered later. As an example, we present a simulated scenario of a network able to detect and cease a wildland fire. It is shown how the network performance depends on its limited resources and how sensors readings and actuators actions depend on each other.