[en] The recent technological developments are boosting the opportunities of accurate method to monitor resource use efficiency in agriculture and in their wake, precision livestock farming (PLF) has experienced huge developments over the past decade. These developments focus on the optimization of individual performances of farm animals as opposed to herd management. The aim of this paper is to explore a method to detect accurately and to analyze changes in cattle's behaviors on pasture during grazing time using signals from the inertial measurement unit (IMU) of mobile devices as a possible tool to manage individual grazing behavior.
Commercial iPhones or iPods, which include a 3-axis accelerometer, a gyroscope and a GPS sensor, are fitted on a halter and placed on the neck of grazing cows. The acquired IMU data are recovered using an open source application (Sensor Data, Wavefrontlabs) and analyzed in a “white-box” model of the cows’ movements. First results using time-domain analysis allowed the detection of grazing behaviors showed accuracies ranging between 84% and 96%, attesting the relevancy of the method. Refined signal processing method will improve the detection but will also inform more about the relative link between the behaviors and the pasture attributes such as sward height, composition and nutritive value.