Technology that improves myoelectric compatibility between a prosthesis and the residual limb has been developed and successfully tested at Aalto University, Helsinki, Finland. Doctoral candidate Dennis Yeung and his research group conducted the study in cooperation with Helsinki University Hospital and Imperial College London.
Prostheses with myoelectric interfaces use machine learning algorithms that help interpret user-generated signals, which can be sensitive to external factors such as sweating and can become weaker over time. Yeung and his team developed a fully automated system that learns during normal use and adapts to varying conditions, reducing the need for adjustments to the prosthesis.
“In this system, the user and the system learn from each other simultaneously. This has potential benefits in improving the convenience and robustness of robotic prostheses,” Yeung said.
The technology was tested in a virtual environment where it was compared to existing systems. To assess upper-limb function and the user interface, a participant with an upper-limb amputation completed Clothespin Relocation Tests.
The results indicated the system can improve the reliability of the prosthetics control systems, thereby reducing financial risk for individuals and public health institutions, the researchers concluded.
Editor’s note: This story was adapted from materials provided by Aalto University.
The open-access study, “Co-adaptive control of bionic limbs via unsupervised adaptation of muscle synergies,” was published in IEEE Transactions on Biomedical Engineering .
