부산대학교 도서관

Objective: We recently proposed a new concept of human-machine interface to control hand prostheses which we dubbed the myokinetic control interface . Such interface detects muscle displacement during contraction by localizing permanent magnets implanted in the residual muscles. So far, we evaluated the feasibility of implanting one magnet per muscle and monitoring its displacement relative to its initial position. However, multiple magnets could actually be implanted in each muscle, as using their relative distance as a measure of muscle contraction could improve the system robustness against environmental disturbances. Methods: Here, we simulated the implant of pairs of magnets in each muscle and we compared the localization accuracy of such system with the one magnet per muscle approach, considering first a planar and then an anatomically appropriate configuration. Such comparison was also performed when simulating different grades of mechanical disturbances applied to the system (i.e., shift of the sensor grid). Results: We found that implanting one magnet per muscle always led to lower localization errors under ideal conditions (i.e., no external disturbances). Differently, when mechanical disturbances were applied, magnet pairs outperformed the single magnet approach, confirming that differential measurements are able to reject common mode disturbances. Conclusion: We identified important factors affecting the choice of the number of magnets to implant in a muscle. Significance: Our results provide important guidelines for the design of disturbance rejection strategies and for the development of the myokinetic control interface, as well as for a whole range of biomedical applications involving magnetic tracking.