3D-DCT Data Aggregation Technique for Regularly Deployed Wireless Sensor Networks

Development of data aggregation techniques is thought as an effective way to save energy in order to prolong the lifetime of wireless sensor networks (WSNs). Particular characteristics of data gathered from spatial-temporal domain may represent certain level of correlation among data values. Based on this observation, we start with analyzing the optimal sampling rate in temporal correlation model to find out the best sleep time of sensor nodes. Then we propose an aggregation technique which exploits the spatial-temporal correlation using a discrete cosine transform (DCT). It transfers the spatial-temporal data into uncorrelated frequency domain coefficients. The WSN is split into several clusters. Original data are aggregated at an aggregation point which acts as a cluster head. The 3D-Zigzag sorting algorithm makes sure that the aggregation point transmits the frequency coefficients from lower frequencies which contain the main energy of the original data to higher frequencies. Simulation results show only a few coefficients are enough to recover original data under high correlation model within the user tolerable distortion rate.