Assembling Sensor Networks

Sometimes sensors need to be assembled in response to an emergency. Such assemblage might take a variety of forms. At one extreme, sensors might converge on a single point. At the opposite extreme, sensors might fill in the position of a lattice. This paper looks at solutions in between these two extremes, in which sensors converge on several different locations. Through analysis and simulations, we evaluate the effectiveness of alternative strategies. We find there is a tradeoff: multiple rally point heuristics converge faster, at the expense of a more fragmented network. We discuss ways of bridging clustered networks. This work has pragmatic implications for emergency responders: in situations of high density, single rally point solutions are superior, and in situations of low density time-to-converge trades off with network integration