Abstract
Introduction: This study is concerned with the control of hand prostheses based on electromyograms. As the prosthesis is moved in space, the socket and electrodes can shift on the residual limb, causing changes in the observed electromyograms. These changes can be misinterpreted by the electronic controller and cause the hand to move unintentionally or fail to execute solicited movements. The goal of this study was to explore the mechanisms related to these myoelectric control failures.
Materials and Methods: To study these phenomena, conventional prosthetic EMG electrodes were augmented with force sensors to record the forces through the devices. These multimodal sensors were then used to control prosthetic hands by 15 users with losses below the elbow. The subjects performed four tasks resembling activities of daily living while the electromyogram signals, force signals, and the performance of the hands (including video images) were recorded. Eight subjects reported a total of 38 control errors, each of which was assigned to one of four failure classes and analyzed.
Results: The article shows examples of the electromyogram and force signals recorded during the control failures and discusses possible causes of the failures. Involuntary opening of the hand was identified as the most common failure, and these failures seemed to be associated with changes in the electrode-skin contact forces. It was not possible to attribute clear causes for 26.2% of the failures.
Conclusions: Knowledge of common failure mechanisms can guide improved design of sockets and electrodes and signal processing to reduce the errors experienced by prosthesis users.
Materials and Methods: To study these phenomena, conventional prosthetic EMG electrodes were augmented with force sensors to record the forces through the devices. These multimodal sensors were then used to control prosthetic hands by 15 users with losses below the elbow. The subjects performed four tasks resembling activities of daily living while the electromyogram signals, force signals, and the performance of the hands (including video images) were recorded. Eight subjects reported a total of 38 control errors, each of which was assigned to one of four failure classes and analyzed.
Results: The article shows examples of the electromyogram and force signals recorded during the control failures and discusses possible causes of the failures. Involuntary opening of the hand was identified as the most common failure, and these failures seemed to be associated with changes in the electrode-skin contact forces. It was not possible to attribute clear causes for 26.2% of the failures.
Conclusions: Knowledge of common failure mechanisms can guide improved design of sockets and electrodes and signal processing to reduce the errors experienced by prosthesis users.
Original language | English |
---|---|
Pages (from-to) | 38-51 |
Number of pages | 14 |
Journal | Journal of Prosthetics and Orthotics |
Volume | 32 |
Issue number | 1 |
Early online date | 16 Dec 2019 |
DOIs | |
Publication status | Published - 1 Jan 2020 |
Keywords
- prosthetic hand
- electromyography
- control failure
- contact force measurement
- biomedical transducers
- multimodal sensors