This paper considers data fusion in ad hoc wireless sensor networks with event detection. Nodes periodically take samples and report their measurements to the fusion center by contention. Two procedures are investigated: network oblivious decision making (NODM) and network aware decision making (NADM). The first procedure makes the decision only when all of the samples are collected, while the second procedure takes into account the network delays; the tradeoff between the decision delay and the random access delay is yielded. Both procedures involve the investigation of the optimal stopping time, but the latter is analyzed as a partially observable Markov decision process.
This is a solid piece of theoretical work with extensive simulations.