Detecting fatigue of sport horses with biomechanical gait features using inertial sensors.
Authors: Darbandi Hamed, Munsters Carolien, Parmentier Jeanne, Havinga Paul
Journal: PloS one
Summary
# Editorial Summary Fatigue detection in sport horses remains clinically valuable for injury prevention and performance optimisation, yet current methods relying on blood lactate and other physiological markers are invasive, require veterinary expertise, and cannot be assessed continuously during work. Darbandi and colleagues deployed body-mounted inertial sensors on 60 sport horses to capture kinematic data during walk and trot before and after both high and low-intensity exercise bouts, extracting biomechanical features from the acceleration and gyroscopic signals to identify fatigue-related changes. Machine learning classification models trained on key fatigue indicators—primarily stance duration, swing duration, and limb range of motion—achieved high diagnostic accuracy for distinguishing fatigued from non-fatigued strides across both gaits. These findings demonstrate that wearable sensor technology can provide objective, real-time, non-invasive fatigue assessment during exercise, potentially enabling field-based monitoring without blood sampling or veterinary attendance. For practitioners, this approach offers a practical pathway toward automated fatigue detection systems that could inform training load management, identify subtle performance declines before clinical lameness emerges, and support evidence-based conditioning programmes.
Read the full abstract on PubMed
Practical Takeaways
- •Wearable inertial sensors can enable real-time fatigue monitoring during training and competition without stopping to draw blood or call a veterinarian.
- •Changes in stance/swing duration and limb range of motion are practical gait markers to watch for during exercise to identify when a horse is becoming fatigued.
- •This technology could help prevent overtraining injuries and optimize conditioning programs by objectively measuring fatigue accumulation during work.
Key Findings
- •Biomechanical features including stance duration, swing duration, and limb range of motion can reliably indicate fatigue in horses during exercise.
- •Machine learning classification models achieved high accuracy for fatigue detection during both walk and trot using inertial sensor data.
- •Non-invasive gait analysis with body-mounted inertial sensors provides an automated, veterinarian-independent method for fatigue detection that does not require invasive sampling.