Classification of cow's behaviors based on 3-DoF accelerations from cow's movements

Abstract

Cow's behavior classification helps people to monitor cow activities, thus the health and physiological periods of cows can be well tracked. To classify the behavior of cows, the data from the 3-axis acceleration sensor mounted on their neck is often used. Data acquisition and preprocessing of sensor data is required in this device. We acquire data from the 3-axis acceleration sensor mounted on the cows'neck and send to the microcontrollter. At the microcontroller, a proposed decision tree is applied in real-time manner to classify four important activities of the cows (standing, lying, feeding, and walking). Finally, the results can be sent to the server through the wireless transmission module. The test results confirm the reliability of the proposed device.