Edge-to-Farm Monitoring of Cattle Oro-Motor Activity Using Ear-Tag Piezoelectric Vibration Sensing and Multi-Scale Machine Learning

This dataset accompanies the manuscript entitled “Edge-to-Farm Monitoring of Cattle Oro-Motor Activity Using Ear-Tag Piezoelectric Vibration Sensing and Multi-Scale Machine Learning.” It includes vibration signals recorded from ear-mounted piezoelectric sensors, machine-learning models, feature extraction scripts, and an embedded deployment pipeline for real-time cattle behavior recognition. The dataset contains: Raw and representative vibration signals (feeding, rumination, unknown behaviors) Trained machine learning models (SVM, XGBoost, and wavelet-based models) Feature extraction and preprocessing scripts (Python) Embedded implementation for ESP32-based real-time inference These resources are provided to ensure reproducibility of the experiments and to support practical deployment in precision livestock monitoring systems.