The Fatigue Assessment Model for Athlete Training by Integrating Inertial Sensors and Multimodal Physiological Signals

Imagine you are trying to figure out how tired a runner is. You could ask them, "How do you feel?" but sometimes people lie, or they just don't know the difference between "tired" and "exhausted." You could check their heart rate, but a racing heart might mean they are running fast and strong, or it might mean they are struggling to keep up. For a long time, coaches had to guess by watching the athlete or relying on a single number, like a heart rate monitor. But what if you could see both the movement and the body's reaction at the exact same time? That is the big question this research tackles: how to build a "super-sense" that combines what an athlete is doing with how their body is feeling, to know exactly when they are pushing too hard.

This study, titled "The Fatigue Assessment Model for Athlete Training by Integrating Inertial Sensors and Multimodal Physiological Signals," is like building a high-tech detective kit for sports. The researchers wanted to solve a puzzle: when an athlete is training, are they doing a great job, or are they just spinning their wheels while getting dangerously tired? To do this, they didn't just look at one thing. They combined two different types of "senses." First, they used inertial sensors (little motion trackers on the wrists) that act like a super-accurate stopwatch and movement camera, recording every shake, turn, and lift. Second, they tracked physiological signals like heart rate, heart rate variability (how much the heart's rhythm changes, which is like a stress meter for the nervous system), and RPE (Rate of Perceived Exertion, which is just the athlete's own score of how hard they feel they are working).

The team put these pieces together to create a model that can tell the difference between "good tired" (working hard and getting better) and "bad tired" (burning out and risking injury). They tested this on a group of athletes doing a 30-minute training session. The results were impressive: the system could recognize exactly what movements the athletes were doing with 97.0% accuracy. More importantly, it could guess their fatigue level with 95.1% accuracy. The data showed that as the training went on, the athletes' heart rates jumped from about 75.00 bpm to 141.34 bpm, and their self-reported tiredness scores (RPE) climbed from 6.32 to 16.90. Meanwhile, their heart rate variability dropped significantly, showing their bodies were under more stress.

The cool part is that this model doesn't just say "you are tired." It tells a story. It can say, "You are moving efficiently, but your heart is working too hard," or "You are tired, but your movements are still perfect." By looking at the combination of motion and body signals, the model found that while the athletes kept their movement efficiency steady (around 57.50% effective training rate), their internal fatigue load kept climbing, with most of the training time falling into a "high load" category. This suggests that even when athletes feel like they are still performing well, their bodies might be screaming for a break.

The researchers argue that relying on just one method—like only asking the athlete how they feel or only checking their heart rate—is like trying to understand a movie by only looking at the sound or only looking at the pictures. You miss the whole story. Their new model proves that by fusing the "what" (the movement) with the "how" (the body's reaction), coaches can get a much clearer, more reliable picture of an athlete's state. It's not a magic cure-all, but it's a powerful tool that suggests we can finally stop guessing and start knowing exactly when to push and when to rest.