Skip to Main Content
Sensor networks (SN) have arisen as one of the most promising monitoring technologies. The recent emergence of small and inexpensive sensors ease the development and proliferation of this kind of networks in a wide range of actual-world applications.1 So far the majority of SN deployments have assumed that sensors can be configured prior to their deployment because the area and events to monitor are well known at design time. Nevertheless, when the purpose of an SN is to monitor the events of an environment such that the distribution and nature of its events is uncertain, we cannot longer assume that sensors can be configured at design time. Instead, sensors must be endowed with the capacity of autonomously reconfiguring and coordinating in order to maximize the amount of information they perceive over time. In this paper, we propose a low cost (in terms of energy and computation) collective distributed algorithm, the so-called collective search diffusion (CDS) algorithm, which allows the sensors in an SN to collaboratively search for the configurations that maximize the information that they perceive based only on their local knowledge. We empirically show that the CDS algorithm helps an SN efficiently monitor environments where various dynamic events occur while showing high degrees of resilience to sensor failures. Both features make the CDS algorithm a suitable tool for monitoring remote and/or hostile uncharted environments.