Data Relevance-Aware Dynamic Sensing Technique with Battery Lifetime Guarantee for Wireless Sensor Nodes

Abstract

Significant effort has been devoted to the development of dynamic monitoring techniques to exploit the temporal correlation among observations with the intention of improving the energy efficiency in wireless sensor nodes. However, the unpredictable nature of these approaches can also be a drawback for their adoption in safety-critical applications. For battery-powered sensor nodes, the predictability of the battery lifetime of the node is paramount for their correct maintenance and operation. Nevertheless, little effort has been made to provide some level of predictability in the available dynamic sampling techniques. The work presented in this article aims to solve this problem by proposing a dynamic monitoring method, capable of exploiting temporal correlation, to improve its energy usage, while at the same time being able to comply with the battery lifetime constraints. The initial evaluation tests show that the proposed sampling method is able to outperform traditional non-adaptive sampling approaches in terms of energy usage, while being able to achieve its battery lifetime goals, even in the presence of disturbances in the energy consumption of the sensor node.