Sign in Subscribe
Machine learning

Artificial Intelligence Enhances Cattle Monitoring on Rugged Rangelands

Strategic Insights

— 1 min read

Researchers at New Mexico State University have successfully deployed artificial intelligence to improve cattle monitoring on large, often rugged, rangelands. Using Long Range Wide Area Network (LoRaWAN) tracking and monitoring technology, the study tested the performance of five machine learning classifiers to discriminate between active and stationary states, and among walking, grazing, ruminating, and resting behaviors of cattle. The research aimed to streamline cattle tracking, which is a daunting task that can be time-consuming and labor-intensive.

Key Takeaways:

  • The study used LoRaWAN cattle collars equipped with a Global Navigation Satellite System (GNSS) receptor and triaxial accelerometer to collect data on cattle behavior.
  • Twenty-four mature cows of four breeds were monitored across four periods between July and November 2022.
  • Behavioral observations were made using 168 hours of video records, resulting in a dataset of 9222 instances of labeled sensor data.
  • Logistic regression, support vector machine, multilayer perceptron, XGBoost, and random forest algorithms were trained and tested, with all classifiers correctly differentiating between states and behaviors with an accuracy and macro-F1 score of 0.95 and 0.90, respectively.
  • The study suggests that the application of trained models can satisfactorily monitor cattle behavior on desert rangeland.
  • The annotated dataset used in this study is publicly available at Perea et al. (1).
  • The research was conducted by Andres Perea, Sajidur Rahman, Huiying Chen, Andrew Cox, Shelemia Nyamuryekung'e, Mehmet Bakir, Huiying Cao, Richard Estell, Brandon Bestelmeyer, Andres F. Cibils, and Santiago A. Utsumi.

Statistics:

  • 24 mature cows of four breeds were monitored across four periods.
  • 38,784 data points were collected using LoRaWAN cattle collars.
  • 9222 instances of labeled sensor data were used to train and test machine learning models.
  • 168 hours of video records were used to observe cattle behavior.
  • The classifiers achieved an accuracy of 0.95 and a macro-F1 score of 0.90.

Sources:

  • Perea et al. [1]
  • Integrating LoRaWAN sensor network and machine learning models to classify beef cattle behavior on arid rangelands of the southwestern United State. Smart Agricultural Technology, 2025,11():101002.
  • The publisher for Smart Agricultural Technology is Elsevier.
  • A free version of this journal article is available at https://doi-org.sdpl.idm.oclc.org/10.1016/j.atech.2025.101002.

Read more