Tradeoff between Delay and Energy Consumption of Partial Data Aggregation in Wireless Sensor Networks

Recently, wireless sensor networks attract researchers’ attention due to its applicability to many fields for effective collection of sensing data within less cost. In wireless sensor networks, due to battery-powered sensor nodes, energy saving is a critical issue. As one of the techniques for energy saving, data aggregation has been proposed. In this paper, we analyze the tradeoff between communication delay and energy consumption of data aggregation. At first, we analyze full aggregation, non-aggregation with Markovian chain. Analytical results show that the non-aggregation method suffers large energy consumption while the full aggregation suffers long transmitting delay. Then, we propose a partial aggregation method called WRP (Waterfalls Random Partial aggregation) which can trade off energy consumption and transmission delay. For the better balance of energy consumption and transmission delay, we utilize several groups of random pushing vectors and investigate them on two criteria based on delay energy products. From the results we find that less arrival data should be aggregated for small data generation rate at nodes to minimize both criteria. For moderate generation rate we must choose a sophisticated random pushing vector depending on the criterion. Furthermore, we extend the analysis in order to discuss the accuracy of sensed data.