Related Experiment Video
Updated: Oct 4, 2025
09:47
Spotting Cheetahs: Identifying Individuals by Their Footprints
Published on: May 1, 2016
15.0K
Categorising cheetah behaviour using tri-axial accelerometer data loggers: a comparison of model resolution and data
Natasha E McGowan1, Nikki J Marks1, Aaron G Maule1
1School of Biological Sciences, Queen's University Belfast, 19 Chlorine Gardens, Belfast, BT9 5DL, UK.
Movement Ecology
|February 6, 2022
Summary
This study used accelerometers to identify fine-scale cheetah behaviors, crucial for conservation. Different devices showed varying accuracy, highlighting the need for careful selection in monitoring vulnerable species.
Area of Science:
- Animal behavior research
- Conservation technology
- Wildlife monitoring
Background:
- Extinction poses a global threat, necessitating effective conservation strategies.
- Understanding species' ecological needs is vital for conservation success.
- Remote sensing technology aids in monitoring animal movement and behavior.
Purpose of the Study:
- To determine fine-scale behaviors in cheetahs (Acinonyx jubatus) using animal-attached accelerometers.
- To compare the performance of different accelerometer devices for behavior categorization.
- To establish a framework for cheetah behavioral classification.
Main Methods:
- Five captive cheetahs were fitted with two types of accelerometer devices (CEFAS and GCDC).
- Behaviors were recorded via video during a lure-chasing activity.
- Accelerometer data were aligned with video footage, and random forest models were used for categorization accuracy assessment at multiple resolutions.
Main Results:
- Fine- and medium-scale behavior categorization achieved 83-88% accuracy.
- The GCDC device generally outperformed the CEFAS device in categorizing non-locomotory behaviors.
- Both devices accurately categorized coarse-scale activity and inactivity, with GCDC showing slightly higher accuracy for activity.
Conclusions:
- The study successfully defined cheetah behaviors beyond basic categories, including accurately identifying stalking via remote sensing.
- Device specifications influence categorization accuracy, suggesting simultaneous deployment of multiple loggers.
- The findings and methods are applicable to monitoring wild cheetahs and other conservation-relevant species.
