Pervasive Sensing for Livestock Health and Activity Monitoring: Current Methods and Techniques

TL;DR

The paper addresses the need for scalable, real-time health and activity monitoring of livestock in large-scale farms. It surveys four pervasive sensing modalities—structural vibration, RF/mmWave, computer vision, and wearables—evaluating their benefits, challenges, and integration potential in smart farming. It finds that vibration sensing offers privacy-preserving, scalable monitoring capable of capturing activity and some vital signs, but faces signal propagation and data-label challenges; CV and RF approaches provide rich behavioral and physiological data yet raise privacy, data management, and environmental robustness concerns, while wearables encounter scalability and durability hurdles. The work highlights gaps in open datasets, data fusion strategies, and deployment-ready solutions, and outlines future directions for interdisciplinary collaboration to realize real-time welfare- and productivity-enhancing smart farming.

Abstract

Pervasive sensing is transforming health and activity monitoring by enabling continuous and automated data collection through advanced sensing modalities. While extensive research has been conducted on human subjects, its application in livestock remains underexplored. In large-scale agriculture, real-time monitoring of biological signals and behavioral patterns can facilitate early disease detection, optimize feeding and breeding strategies, and ensure compliance with welfare standards. This survey examines key sensing technologies -- including structural vibration, radio frequency (RF), computer vision, and wearables -- highlighting their benefits and challenges in livestock monitoring. By comparing these approaches, we provide insights into their effectiveness, limitations, and potential for integration into modern smart farming systems. Finally, we discuss research gaps and future directions to advance pervasive sensing in livestock health and activity monitoring.