A Two-Prong Approach to Energy-Efficient WSNs: Wake-Up Receivers plus Dedicated, Model-Based Sensing

Abstract

Energy neutral operation of WSNs can be achieved by exploiting the idleness of the workload to bring the average power consumption of each node below the harvesting power available. This paper proposes a combination of state-of-the-art low-power design techniques to minimize the local and global impact of the two main activities of each node: sampling and communication. Dynamic power management is adopted to exploit low-power modes during idle periods, while asynchronous wake-up and prediction-based data collection are used to opportunistically activate hardware components and network nodes only when they are strictly required. Furthermore, the concept of "model-based sensing" is introduced to push prediction-based data collection techniques as close as possible to the sensing elements. The results achieved on representative real-world WSN case studies show that the combined benefits of the design techniques adopted is more than linear, providing an overall power reduction of more than 3 orders of magnitude.