Smart Livestock Monitoring Using IoT and Biosensor Technologies

Abstract

Real-time monitoring of livestock health, behaviour, and physiological status is critical for early disease detection,
reproductive management, and productivity optimisation in modern animal production systems. This study presents the
design, deployment, and validation of an integrated Internet of Things (IoT) and biosensor platform for continuous smart
livestock monitoring across cattle, sheep, and pig production units in Estonia, Austria, and Switzerland. The system
architecture comprised RFID ear-tag accelerometers for activity and rumination monitoring, subcutaneous
microchip-linked temperature loggers, rumen pH boluses, collar-mounted heart-rate sensors, and LoRaWAN-connected
edge gateways transmitting to a cloud analytics platform. A total of 847 animals were monitored over 18 months (January
2024-June 2025). Machine learning classifiers--including gradient boosting and LSTM networks--were trained on sensor
fusion data to detect oestrus, fever, lameness, and subacute ruminal acidosis (SARA) events. The IoT-ML system
achieved sensitivity of 94.3% and specificity of 91.7% for fever detection, 89.6% and 93.2% for oestrus detection in
cattle, and 87.4% and 90.1% for SARA identification. Compared to conventional twice-daily visual observation, the
platform reduced disease event detection latency by 73% and increased oestrus detection rate by 31%. These results
demonstrate that IoT-biosensor integration with edge-to-cloud ML analytics constitutes a scalable, cost-effective solution
for precision livestock farming across diverse European production environments.