Stationary Wavelet Processing and Data Imputing in Myoelectric Pattern Recognition on a Low-Cost Embedded System
A challenge when analysing electric signals from the muscles using artificial intelligence algorithms to control prosthetic limbs is to account for signal noise and faulty sensors. There is a class of processing that are wavelet based which usually cannot be used in real-time that can mitigate signal noise and faulty sensors. We tested wavelet-based processing on nine able-bodied participants with inconclusive results and on a dataset from 15 participants which showed improvements. These results suggest that wavelet-based processing can be used in real-time, but more research is needed to determine the trade-off between robustness and responsiveness.