Entropy correlation-based clustering method for representative data aggregation in wireless sensor networks

One of the popular data aggregation method in wireless sensor network (WSN) is collecting only local representative data based on correlation of sample data. To recognise the local representative nodes, it is necessary to determine the correlation regions. However, recent correlation models are distance based that is not general and need to be determined beforehand or complicated with high computing cost. Thus, in this paper, a novel entropy correlation model is proposed based on joint entropy approximation. Using the proposed model, an entropy correlation-based clustering method is presented and the selection of representative data that satisfying the desired distortion is proposed. The algorithm is validated with practical data.