Sensor networks: decentralized monitoring and subspace classification of events

Abstract

Sensor networks are an emerging technology that promises fast and easy monitoring of the physical world. One of the applications of sensor networks is environmental monitoring, which consists of a large number of sensors collecting massive data about the atmosphere, thus making centralized monitoring extremely difficult. In this paper, we show that centralized monitoring is not necessary due to the fact that there exist different types of events in the data, each of which can be effectively monitored by a small subset of sensors. We propose decentralized monitoring of a sensor network that automatically identifies these event types and the related groups of sensors. Our contribution is twofold: (1) the proposed decentralized solution achieves event classification performance equivalent to the centralized solution and (2) it mines valuable information (event types and groups of sensors) useful for researchers studying events and sensor deployment strategies. We provide a thorough evaluation of the proposed solution, conduct extensive experiments using both benchmark and real-world sensor data, and observe consistent performance. We suggest some further work based on our study and experiments.