Delay Tolerant Dynamic Data Collection over a Sensor Network

The work presented in this paper consists in experimenting on the collection of dynamic data distributed over a sensor network, where only neighborhood communication is supported (no routing) and parts of the network can become disconnected (no flooding). This can be used to track physical phenomena in the context of battlefields, crisis management or disaster response. By observing the evolution of the distributed variables that make the phenomenon, it is possible to work out emerging trends and potential threats, and to undertake the required actions, i.e. to support situation management. Our goal in this paper is to describe the framework that we have designed and implemented and that makes it possible to cope with missing measurements. When some data items are missing, either because of interferences or faulty sensors, it still enables to provide a continuous view of the distributed phenomenon by using interpolation or extrapolation techniques. Our focus is on the framework to support the implementation of such features, not on the final applications. Nevertheless we illustrate its use it by means of a temperature collection application.