Related Experiment Video
Updated: Jul 20, 2026
07:52
Automated Rat Single-Pellet Reaching with 3-Dimensional Reconstruction of Paw and Digit Trajectories
Published on: July 10, 2019
14.3K
Evaluating Behavior Recognition Pipeline of Laying Hens Using Wearable Inertial Sensors.
Kaori Fujinami1,2, Ryo Takuno2, Itsufumi Sato3
1Division of Advanced Information Technology and Computer Science, Institute of Engineering, Tokyo University of Agriculture and Technology, Tokyo 184-8588, Japan.
Sensors (Basel, Switzerland)
|June 10, 2023
Summary
This study introduces a wearable sensor system for monitoring hen behavior, enhancing animal welfare in poultry farming. Machine learning accurately recognizes 12 distinct behaviors, improving welfare-oriented rearing systems.
Area of Science:
- Animal Science
- Biotechnology
- Machine Learning
Background:
- Growing global concern for animal welfare necessitates improved farming practices.
- Conventional battery cage systems for laying hens may compromise their physical and mental well-being.
- Welfare-oriented rearing systems are being developed to balance animal welfare with agricultural productivity.
Purpose of the Study:
- To develop and evaluate a behavior recognition system using wearable inertial sensors for continuous monitoring of laying hen behaviors.
- To improve welfare-oriented rearing systems through objective quantification of animal behavior.
- To identify optimal system parameters for accurate behavior recognition.
Main Methods:
- A supervised machine learning approach was employed to recognize 12 distinct hen behaviors.
- Wearable inertial sensors (accelerometer and angular velocity) were utilized for data collection.
- Parameters such as classifier type (multi-layer perceptron), sampling frequency (100 Hz), window length (1.28 s), and data imbalance handling were investigated.
Main Results:
- The system successfully recognized a variety of 12 hen behaviors with a reference configuration.
- The study analyzed the impact of different parameters on recognition accuracy.
- The findings provide a basis for optimizing similar behavior recognition systems.
Conclusions:
- Wearable inertial sensors and machine learning offer a viable method for continuous monitoring and quantification of hen behaviors.
- This technology can significantly contribute to enhancing animal welfare in poultry production.
- The results facilitate the design of advanced welfare-monitoring systems and inform parameter selection for specific applications.
