A novel compressive sensing based Data Aggregation Scheme for Wireless Sensor Networks

The random distribution of sensors and the irregularity of routing paths lead to unordered sensory data which are difficult to deal with in Wireless Sensor Networks (WSNs). However, for simplicity, most existing researches ignore those characteristics in the designs of Compressive Sensing based Data Aggregation Schemes (CSDAS). Since conventional sparsification bases (e.g., DCT, Wavelets) are inefficient to deal with unordered data, performances of CSDAS with conventional bases are inevitably constrained. In this work, a novel CSDAS which adopts Treelet transform as a sparse transformation tool is proposed. Our CSDAS is capable to exploit both spatial relevance and temporal smoothness of sensory data. Moreover, our CSDAS contains a novel correlation based clustering strategy which is realized with the localized correlation structure of sensory data returned by Treelets and facilitates energy saving of CSDAS in WSNs. Comparative results show the reconstruction error rate with adopting Treelet transform in CSDAS is about 18% lower than that of conventional ones when the normalized energy consumption is 0.3. Even larger performance gain will be obtained at higher energy consumption level. Meanwhile, simulations results further show that our novel correlation based clustering strategy is of great potential. Specially, there is a gain of roughly 35% for total energy savings with our proposed clustering strategy.