Data Sweeper: A Proactive Filtering Framework for Error-Bounded Sensor Data Collection

Abstract

This paper presents data sweeper—a novel framework that attempts to reduce network traffic for error-bounded data collection in wireless sensor networks. Unlike existing passive filters, a data sweeper migrates in the network and proactively suppresses data updates while maintaining the user-defined error bound. Intuitively, the migration of a data sweeper learns the data change of each sensor node on the fly, which helps to maximize the filtering capacity. We design the data sweeper framework in such a way that it can accommodate diverse query specifications and be easily incorporated into the existing sensor network protocols. Moreover, we develop efficient strategies for query precision maintenance, sweeper migration, and data suppression within the framework. In particular, in order to maximize traffic reduction and adapt to online data updates, a Lagrangian relaxation-based algorithm is proposed for data suppression. Extensive simulations based on real-world traces show that the data sweeper significantly reduces the network traffic and extends the system lifetime under various network configurations.