Case study · sEMG ankle-movement decoding
Reading Ankle Movement From Muscle Signals
Surface EMG decodes ankle movements in stroke patients: time/frequency features, boosted trees.
Decoding ankle-joint movements in stroke patients from surface electromyography (sEMG), part of a master's thesis where I provided technical support and carried out the biomedical signal-processing and machine-learning work in MATLAB. Time- and frequency-domain features, ANOVA feature selection and a boosted-trees classifier reached 94% accuracy.
- Org
- MS thesis
- Status
- Research
- Channels
- CH2CH3
The problem
Stroke rehabilitation needs to know how a patient's ankle is moving. Can surface EMG from the leg muscles decode ankle-joint movements reliably?
What I built
- Part of a master's thesis: I provided technical support and carried out the signal-processing and ML work
- Time- and frequency-domain features from surface EMG
- ANOVA feature selection and a boosted-trees classifier
Stack
MATLAB · sEMG · ANOVA · Boosted trees
Results
- 94% ankle-movement decoding accuracy
Surface EMG electrodes on the lower-leg muscles capture the electrical activity behind each ankle movement. The pipeline extracts time-domain and frequency-domain features, ranks them with ANOVA, and trains a boosted decision-tree ensemble to recognise the movement.
Reliable movement decoding from muscle signals is a building block for stroke rehabilitation tools, assistive devices and EMG-controlled wearables.