Animal welfare monitoring relies on sensor accuracy for detecting changes in animal well-being. We compared the distance calculations based on global positioning system (GPS) data alone or combined with motion data from triaxial accelerometers. The assessment involved static trackers placed outdoors or indoors vs. trackers mounted on cows grazing on pasture. Trackers communicated motion data at 1 min intervals and GPS positions at 15 min intervals for seven days. Daily distance walked was determined using the following: (1) raw GPS data (RawDist), (2) data with erroneous GPS locations removed (CorrectedDist), or (3) data with erroneous GPS locations removed, combined with the exclusion of GPS data associated with no motion reading (CorrectedDist_Act). Distances were analyzed via one-way ANOVA to compare the effects of tracker placement (Indoor, Outdoor, or Animal). No difference was detected between the tracker placement for RawDist. The computation of CorrectedDist differed between the tracker placements. However, due to the random error of GPS measurements, CorrectedDist for Indoor static trackers differed from zero. The walking distance calculated by CorrectedDist_Act differed between the tracker placements, with distances for static trackers not differing from zero. The fusion of GPS and accelerometer data better detected animal welfare implications related to immobility in grazing cattle.

Virtual fencing systems have emerged as a promising technology for managing the distribution of livestock in extensive grazing environments. This study provides comprehensive documentation of the learning process involving two conditional behavioral mechanisms and the documentation of efficient, effective, and safe animal training for virtual fence applications on nursing Brangus cows. Two hypotheses were examined: (1) animals would learn to avoid restricted zones by increasing their use of containment zones within a virtual fence polygon, and (2) animals would progressively receive fewer audio-electric cues over time and increasingly rely on auditory cues for behavioral modification. Data from GPS coordinates, behavioral metrics derived from the collar data, and cueing events were analyzed to evaluate these hypotheses. The results supported hypothesis 1, revealing that virtual fence activation significantly increased the time spent in containment zones and reduced time in restricted zones compared to when the virtual fence was deactivated. Concurrently, behavioral metrics mirrored these findings, with cows adjusting their daily travel distances, exploration area, and cumulative activity counts in response to the allocation of areas with different virtual fence configurations. Hypothesis 2 was also supported by the results, with a decrease in cueing events over time and increased reliance with animals on audio cueing to avert receiving the mild electric pulse. These outcomes underscore the rapid learning capabilities of groups of nursing cows in responding to virtual fence boundaries.

This study classified cows’ foraging behaviors using machine learning (ML) models evaluated through random test split (RTS) and cross-validation (CV) data partition methods. Models included Perceptron, Logistic Regression, Support Vector Machine, K-Nearest Neighbors, Random Forest (RF), and XGBoost (XGB). These models classified activity states (active vs. static), foraging behaviors (grazing (GR), resting (RE), walking (W), ruminating (RU)), posture states (standing up (SU) vs. lying down (LD)), and posture combinations with rumination and resting behaviors (RU_SU, RU_LD, RE_SU, and RE_LD). XGB achieved the highest accuracy for state classification (74.5% RTS, 74.2% CV) and foraging behavior (69.4% CV). RF outperformed XGB in other classifications, including GR, RE, and RU (62.9% CV vs. 56.4% RTS), posture (83.9% CV vs. 79.4% RTS), and behaviors-by-posture (58.8% CV vs. 56.4% RTS). Key predictors varied: speed and Actindex were crucial for GR and W when increasing and for RE and RU when decreasing. X low values were linked to RE_SU and RU_SU, while X and Z influenced RE_LD more. RTS showed higher accuracy in activity states classification while CV in foraging behaviors and by posture classification. These results emphasize CV in RF’s reliability in managing complex behavioral patterns and the importance of continuous recording devices and movement data to monitor cattle behavior accurately.

Body condition score (BCS) has been a useful tool in estimating the health of cattle for many years now. This categorical metric requires experienced observers to visually inspect cows and assess body fat deposits regularly via a time consuming, subjective process. Low cost RGB+depth cameras have been used alongside machine learning algorithms in the past and have shown great promise, however, more advanced techniques are projected to yield better performance. In this work, a vision transformer (ViT) is pretrained using a recently developed self-supervised pretraining method, masked image modeling, and then fine-tuned on RGB+depth BCS data with the objective of improving performance. Model accuracy was found to be highly dependent on dataset curation, ranging from 64% to 92% accuracy. These discrepancies are attributed to non-unique data in the training and test splits and an inherently unbalanced dataset, both of which are discussed in detail. It is recommended that engineers and animal scientists collaborate more closely, as certain details related to dataset curation are critical to thoroughly assess performance and robustness of automated methods for BCS determination.