Table 7.
Summary of characteristic of studies that investigated muscle fatigue quantification using wearable devices.
| No. | Type of study | # of participants | Fatiguing task | Task duration | Input data | Reference measure | Modeling approach | Output | Model performance |
|---|---|---|---|---|---|---|---|---|---|
| 57 | Lab | 24 | 2-min squatting at 8 squats/min | Time until Borg's RPE ≥17 | Motion | Borg's RPE | Support vector machine | Binary/4 levels | ACC = 91/61% |
| 58 | Lab | 10 | Shoulder flexion, shoulder abduction, elbow extension performed using a Barrett WAM arm | Time until self-reported fatigue + 10 s (3 repetitions per exercise) | EMG | Self-report | Gradient Boosting | Binary |
F1-score = 70.4–76.6% Success rate = 73–74% |
| 59 | Lab | 3 | Muscle fatiguing exercise | Time until self-reported fatigue | EMG | Not stated | Backpropagation neural networks | Binary | ACC = 100% |
| 60 | Field | 5 | Isotonic/isometric bicep curl, Isotonic/isometric leg extension | Time until failure (1–2 min) | Motion | Gradient of the Dimitrov spectral index (based on EMG) | Regression tree model | Gradient of the relative change in the Dimitrov spectral fatigue index | Error <15% |
EMG,electromyogram; RPE,ratings of perceived exertion; ACC,accuracy.