Towards a Taxonomy of Task Allocation in Sensor Networks

A sensor network often consists of a large number of sensing devices of different types. These can vary from very simple sensors with limited capabilities such as motes, to very complex systems such as Unmanned Aerial Vehicles (UAV). Upon deployment in the field, these different sensors form an ad hoc network using wireless links or cables to communicate with each other and with data processing centres. Given that sensing devices can be static or mobile, and that sensors may fail or be damaged, it is clear that the network configuration must be highly dynamic. An already deployed sensor network is usually required to support multiple sensing tasks of different types to be accomplished simultaneously. Sensing tasks might share the usage of a sensing resource, but more often they compete to exclusively control it. Consider for example the search-and-rescue scenario of Figure 1, where the sensor network has to support an identification task (of people in need of rescue), and at the same time a wide area surveillance task (to detect possible threats to the life of the injured people). If only one UAV is available to be allocated to a task, the question to answer is: “Where is it better to send that particular UAV?”.