Advancing Cattle Welfare: Ultra Low-power Health Monitoring at the Edge.

Cattle health problems, such as Bovine Respiratory Disease (BRD), result in significant economic losses in agriculture industry. The current management practices to diagnose and select cattle for treatment is a widespread clinical scoring system called DART (Depression, Appetite, Respiration, and Temperature). DART requires significant manual human labor since animal evaluation is done individually. We propose a novel IoT system that utilizes wearable accelerometers to predict DART scores, reducing labor and costs associated with manual evaluation. The proposed system processes accelerometer data to capture cattle behavior and uses a lightweight decision-tree classifier for prediction. Evaluation on a dataset of 54 animals shows 78% accuracy in classifying healthy and sick animals. Compared to 13 state-of-the-art classifiers, our approach outperforms them in accuracy and computational complexity. With low resource usage and energy consumption, our IoT solution is suitable for deployment in smart farms.