Energy-Efficient Edge–Cloud IoT-Based Architecture for Automated Hive Monitoring of Pollinator Species in Precision Agriculture

Abstract

Pollinator insect species play a vital role in crop pollination and precision agriculture. Certain insect species are facing a decline due to the excessive use of fertilizers and climate change, putting global food security at risk and necessitating continuous automated surveillance. Existing surveillance systems suffer from high power consumption due to the processing of large volumes of captured data on resource-constrained hardware. To address these challenges, this study presents a sensing system that overcomes the inefficiencies of traditional detection frameworks through a proposed on-device, energy-efficient architecture. The designed architecture integrates a customized framework that streamlines the processing pipeline, thereby reducing computational load, on-device energy usage, and transmission-related power consumption. Experimental results demonstrate that the proposed architecture significantly reduces energy consumption compared to traditional frameworks. These reductions translate into a substantial improvement in battery life, ranging from 38% to 107% across different operational conditions. Moreover, to enable remote surveillance, stable Internet connectivity is required, which is unfeasible in rural regions. Therefore, the proposed sensing system enables remote optimized data extraction features without disrupting ongoing processes, ensuring seamless functionality in both high and low internet connectivity regions by utilizing 4G cloud-based transmission in the former and local Wi-Fi hotspot access in the latter. Overall, the proposed system outperforms existing systems with significantly lower current consumption and extended battery lifetime.