Sign in Subscribe
Agriculture

AI-Powered Sheep Behavior Classification for Precision Farming

Strategic Insights

— 1 min read

Artificial intelligence is revolutionizing the way we farm, especially in precision livestock farming. A recent study from Telkom University in Indonesia has developed an efficient and real-time animal behavior monitoring system using a triaxial accelerometer and embedded machine learning models. This innovative approach aims to support precision farming by classifying sheep behaviors such as standing, laying, and sleeping with high accuracy. The system has been successfully deployed on a low-power IoT device and has shown promising results in a real-world setting.

Key Takeaways:

  • The research proposes a behavior classification system for sheep using a triaxial accelerometer and embedded machine learning models implemented on a low-power IoT device.
  • The study evaluates four classification algorithms: Random Forest, K-Nearest Neighbors, Support Vector Machine, and XGBoost, and finds that Random Forest outperforms other models with an accuracy of 92.1% and an F1-score of 0.9272.
  • The system reduces the reliance on manual observation and enables scalable implementation in practical farm environments.
  • The authors suggest incorporating additional sensor modalities, expanding behavior classes, and adopting adaptive learning techniques to improve robustness in diverse and dynamic livestock settings.
  • The system demonstrates that lightweight machine learning models can be effectively integrated into embedded platforms for autonomous animal behavior monitoring.
  • Al Fahri Suhaimi, Giva Andriana Mutiara, and Muhammad Fahru Rizal are the researchers involved in this study.
  • The study was funded by the Research And Community Service Directorate (Ppm) of Telkom University.

Statistics:

  • The system achieves an accuracy of 92.1% in classifying sheep behaviors.
  • The F1-score of the system is 0.9272.
  • The system transmits classified data via Wi-Fi with a latency under 2 seconds.
  • The system demonstrates a reduction in manual observation which enables scalable implementation in practical farm environments.
  • 6,000 labeled samples were used in the experimental results.
  • The study proposes a behavior classification system for sheep behaviors such as standing, laying, and sleeping.

Sources:

  • Leveraging Embedded Accelerometers and Machine Learning for Real-Time Sheep Behavior Classification in Precision Farming. IEEE Access, 2025, 13():165006-165024. (IEEE Access - http://ieeexplore.ieee.org/servlet/opac?punumber=6287639)
  • Telkom University. Telkom University Researchers Describe New Findings in Machine Learning (Leveraging Embedded Accelerometers and Machine Learning for Real-Time Sheep Behavior Classification in Precision Farming). Journal of Engineering. October 13, 2025; p 5308.

Read more