Classification of individual dairy cow behaviors using accelerometer, gyroscope, and integrated sensor models.

BACKGROUND: Automated behavior monitoring is increasingly important in precision dairy farming, supporting early disease detection, welfare assessment, and productivity optimization. Although accelerometers effectively detect postural changes, they have limited capacity to capture rotational or transitional movements. Gyroscopes provide complementary angular velocity data that may enhance classification of complex behaviors. However, their combined use remains underexplored, particularly at the individual cow level. This study aims to evaluate the performance of accelerometer, gyroscope, and combined sensors models for classifying four key cow behaviors: lying, standing, eating, and walking at the individual animal level. RESULTS: Over 780,000 labeled observations were obtained from seven dairy cows monitored over 90 days. Lying behavior consistently produced low, stable signals across all axes of accelerometer and gyroscope, while eating showed the greatest variability, particularly along the X and Y axes. Significant axis-specific and behavior-specific differences were observed (p < 0.05), with GyroY and GyroZ capturing the highest rotational activity during eating and walking. Signal vector magnitudes effectively distinguished behaviors, with lying showing the lowest values and eating the highest. Random Forest models combining accelerometer and gyroscope data consistently outperformed single-sensor approaches, particularly for classifying lying and standing behaviors. Although eating and walking exhibited lower sensitivity, sensor fusion improved classification robustness across individuals. CONCLUSION: The integration of accelerometer and gyroscope data enhanced classification accuracy, particularly for static behaviors. Axis-specific signal patterns and individualized modeling revealed critical insights into behavior differentiation and cow-specific variability. These findings support the development of scalable, sensor-based monitoring systems tailored to precision livestock management.