Dairy cattle monitoring using wireless acceleration-sensor networks

We present novel methods for detecting estrus and predicting calving-time in dairy cattle, by using a wireless network with an acceleration sensor. Our estrus detection method focuses on the habit of cows in estrus to become restless. A simple index - the hourly durations when the minute-based standard deviation of the acceleration sensor values for a leg is high - was experimentally proved to be effective at recognizing estrus. Our method of calving-time prediction utilizes the phenomenon that a sensor module placed in the vagina of a preparturient cow is expelled by the rupture of the membrane, which usually occurs a couple of hours prior to parturition. The phenomenon can be easily detected by focusing on the fluctuation in attitude of the sensor module and/or the change in session state of the wireless network before and after membrane rupture. Thus, upcoming parturition can be recognized.