Improved prosthetic control based on myoelectric pattern recognition via wavelet-based de-noising

The use of myoelectric pattern recognition (MPR) for the control of prosthetic limbs has been limited by interfering noise and susceptibility to motion artifacts. In this article, we present a novel algorithm using wavelet transforms that can be executed in real-time and improves the robustness of MPR systems. The algorithm outperformed conventional methods and showed potential for improving the feasibility and usability of prosthetic devices in real-life situations.