Energy-efficient data aggregation and transfer in periodic sensor networks

Limited battery power and high transmission energy consumption in wireless sensor networks make in-network aggregation and prediction a challenging area for researchers. The most energy consumable operation is transmitting data by a sensor node, comparing it with the energy consumption of in-network computation which is negligible. The energy trade-off between communication and computation provides applications benefit when processing the data at the network side rather than simply transmitting sensor data. In this study, the authors consider a cluster-based technique with which data is sent periodically from sensor nodes to their appropriate cluster-heads (CH). The proposed technique manages energy efficiency in periodic sensor network and it consists of two phases: `aggregation phase and adaptation phase´. The aggregation phase is used to find similarities between data (measurements captured during a period p) in order to eliminate redundancy from raw data, thus reducing the amount of data-sets sent to the CH. The adaptation phase provides sensors the ability to identify duplicate data-sets captured among successive periods, using the sets-similarity joins functions. To evaluate the performance of the proposed technique, experiments on real sensor data have been conducted. Results show that the proposed technique is effective in term of energy consumption and quality of data.