Analysis of Accelerometer and GPS Data for Cattle Behaviour Identification and Anomalous Events Detection

Abstract

In this paper, a method to classify behavioural patterns of cattle on farms is presented. Animals were equipped with low-cost 3-D accelerometers and GPS sensors, embedded in a commercial device attached to the neck. Accelerometer signals were sampled at 10 Hz, and data from each axis was independently processed to extract 108 features in the time and frequency domains. A total of 238 activity patterns, corresponding to four different classes (grazing, ruminating, laying and steady standing), with duration ranging from few seconds to several minutes, were recorded on video and matched to accelerometer raw data to train a random forest machine learning classifier. GPS location was sampled every 5 min, to reduce battery consumption, and analysed via the k-medoids unsupervised machine learning algorithm to track location and spatial scatter of herds. Results indicate good accuracy for classification from accelerometer records, with best accuracy (0.93) for grazing. The complementary application of both methods to monitor activities of interest, such as sustainable pasture consumption in small and mid-size farms, and to detect anomalous events is also explored. Results encourage replicating the experiment in other farms, to consolidate the proposed strategy.

Cow Behavioural Activities in Extensive Farms: Challenges of Adopting Automatic Monitoring Systems

“…In Cabezas et al [49], the animals wore a low-cost commercial device equipped with 3D accelerometers and GPS sensors. The accelerometer sampling frequency used was 10 Hz and the GPS acquisition interval was set to 5 min.…”

Behavioral Fingerprinting: Acceleration Sensors for Identifying Changes in Livestock Health

“…In order to accurately classify animal activity, some other 3-axis acceleration-related sensors that may not be included in Table 1, have been also used or developed, such as Silent Herdsman [122,123], ProMove-mini [124], iFarmTec [125], MPU9250 9-axis microelectromechanical system [73], MinIMU-9V2 IMU [126], Digitanimal Livestock GPS [127], GPS collar [128], Bosch BMI160 [75,129] and Bosch BMA400 micro electromechanical system [130]. Further, these sensors are utilized in combination with additional sensors or/and approaches of data processing and analysis for predicting animal behaviors.…”

“…The Digitanimal Livestock GPS device is integrated with a 3-D micro-electromechanical-system accelerometer and a GPS sensor, which was attached to the neck of cattle to obtain the accelerometer raw data at a sampling frequency of 10 Hz together with video recording on the durations of grazing, ruminating, laying and steady standing, and a random forest machine learning algorithm was used to classify cattle behaviors matched to accelerometer records with good accuracies of 0.93, 0.907. 0.881, and 0.922 for grazing, ruminating, laying and steady standing, respectively [127]. A lab-constructed GPS collar, which is comprised of an iGotU GT-120 GPS logger and a 3-axis X16 mini accelerometer, was mounted to the bottom of the cattle's neck to classify grazing and non-grazing behaviors using random forest (RF), linear discriminant analysis (LDA), quadratic discriminate analysis (QDA), and support vector machines (SVM) for comparison [128].…”

Assessment of a UWB Real Time Location System for Dairy Cows’ Monitoring

“…In detail, collar sensors and leg sensors are able to monitor feeding and ruminating, and lying and standing, respectively. To increase information on animal behaviour and overcome this limitation, a combination of different systems is applied in both indoor (i.e., dairy houses, cattle houses) [ 21 , 28 ] and outdoor (i.e., grazing) [ 29 , 30 ]. One of the positive aspects is that the identification of the animal’s position by a combination of sensors could provide hints about animal behaviours.…”