Skip to Main Content
Wireless sensor networks (WSNs) have a broad range of applications, such as battlefield surveillance, environmental monitoring, and disaster relief. These networks usually have stringent constraints on the system resources, making data-extraction and aggregation techniques critically important. However, accurate data extraction and aggregation is difficult, due to significant variations in sensor readings and frequent link and node failures. To address these challenges, we propose data-aggregation techniques based on statistical information extraction that capture the effects of aggregation over different scales. We also design, in this paper, an accurate estimation of the distribution parameters of sensory data using the expectation-maximization (EM) algorithm. We demonstrate that the proposed techniques not only greatly reduce the communication cost but also retain valuable statistical information that is otherwise lost in many existing data-aggregation approaches for sensor networks. Moreover, simulation results show that the proposed techniques are robust against link and node failures and perform consistently well in broad scenarios with various network configurations.