A Novel Data Fusion Method in Wireless Multimedia Sensor Networks

Multimedia data fusion in wireless sensor networks has attracted great attention in the literature. However the existing data fusion does not take into account the cooperative redundancy of audio data. In order to improve energy efficiency and extend network lifetime, the paper presents an integrated audio data fusion framework, which combines wavelet transform with seamless audio-editing techniques, for cooperative data fusion. The basic idea is to select audio data from a sensor node as the template and transform the appearance of the given audio data toward the selected audio data. Our fusion approach decomposes audio data into mulitple parts. For the different parts, we utilize a graphmatching algorithm to obtain the most appropriate template and transform the shape information of the graph with the obtained template, while for the given audio data, we compare them with the templates and transmit the mismatch parts. Through theoretical justifications and empirical studies, we demonstrate that the proposed scheme achieves substantially superior performances over conventional methods in terms of energy efficiency and data transmission under the resourceconstrained wireless multimedia sensor networks. Keywords–wireless multimedia sensor networks; data fusion; audio data; wavelet transform;