An Energy-Saving Data Statistics-Driven Management Technique for Bio-Powered Indoor Wireless Sensor Nodes

Abstract

Wireless sensor networks (WSNs) are technologies that play an important role in Internet of Things (IoT) systems. In most applications, sensor nodes are expected to operate for extended periods without maintenance. Therefore, minimizing power consumption and self-sustainability based on energy harvesting (EH) still represent important research challenges. In this article, a power management strategy (PMS) based on the weighted order statistics (WOS) classification technique is proposed to dynamically adapt the duty cycle of the sensor node according to historical data measurements. Unnecessary acquisition and transmission of slow-varying signals are reduced, improving the power consumption of the node. As traditional renewable sources (sunlight, wind, and vibration) are scarce in indoor scenarios, an array of Dypsis Lutescens plants is used as a power bioenergy cell, providing a clean and cost-effective alternative to power indoor sensor nodes. The WOS technique is programmed into an nRF52840 microcontroller, and an ultralow-power BQ25570 harvesting circuit harnesses the energy from the array of plants. Experimental results include the energy performance analysis of the wireless sensor node and the measured power generation capacity of the bioenergy source, showing an energy-autonomous behavior for IoT applications. Average power consumption of 15.6 mW per transmitted data packet of 14 bytes is achieved, which represents the ability to perform at least 693 data transmissions per day considering an 8 F supercapacitor as a storage device.