Modeling Wireless Sensor Network with Spatial Constrained Affinity Propagation

Information processing in wireless sensor networks, demands a theory of computation that allows the minimization of processing effort, and the distribution of this effort is relative to the whole network. When monitoring spatial phenomena, such as temperature or humidity in an indoor environment, it is obvious that some sensors might represent their neighbors well with permitted error. Recently, a new clustering algorithm, named ”affinity propagation”, is proposed. Different from the popular k-centers clustering technique, affinity propagation operates by simultaneously considering all data points as potential cluster centers (called ”exemplars”) and exchanging messages between data points until a good set of exemplars and cluster emerges. In this paper, we apply affinity propagation for choosing exemplars in wireless sensing network. However, a difference is made between the original form AP and the algorithm in our application. The AP algorithm in our experiments is exploited thoroughly, under different spatial constraints. We only consider the relationship between close neighbor nodes under a certain threshold of distance, instead of the pairwise similarities between the whole network nodes. The experiments proved that our methodology can also effectively acquire necessary information of network status. Meanwhile, the scarce resources in network (energy, etc.) can be saved in a more efficient way.