AviEar: An IoT-Based Low-Power Solution for Acoustic Monitoring of Avian Species

Abstract

Birds play a pivotal role in maintaining global biodiversity by serving as vital agents in the key ecosystem functions, such as seed dispersal, insect regulation, and pollination. However, escalating anthropogenic pressures such as deforestation, poaching, and climate change have increasingly imperiled the avian populations worldwide. Consequently, effective monitoring strategies are essential for conservation efforts. However, traditional monitoring methods often fall short due to limitations in power efficiency and data storage. This article presents the design and development of AviEar, a novel wireless sensor node tailored for monitoring of avian species. The developed node is the Internet of Things (IoT) device that hosts a MEMS microphone, an ultralow-power advanced RISC machine (ARM) Cortex microcontroller unit (MCU), and a storage unit. The proposed system seamlessly integrates acoustic data recording, on-board signal processing, storage, and cloud-based uploads to facilitate remote monitoring. A standout feature is its rapid target species detection algorithm (DA), approximately executing within a mere 1.443 s. Without real-time onboard processing, the system would generate redundant data and experience increased battery drain. Its real-time selective logging and transmission framework yields an impressive operational span of up to two months at an 8-kHz sampling rate. The field experiments demonstrate AviEar’s ability to provide avian acoustic data with 99.6% precision, 95% recall, 97.2% ${F}1$ -score, a mean 0.77 confidence score, and remarkable power efficiency, showcasing its suitability for sustainable monitoring solutions. Moreover, the outcomes of these deployments furnish conservation decision-makers and researchers with invaluable datasets, empowering them to conduct comprehensive and large-scale monitoring initiatives.