Continuous Proportional Myoelectric Control of an Experimental Powered Lower Limb Prosthesis During Walking Using Residual Muscles.

摘要

Current robotic lower limb prostheses rely on intrinsic sensing and finite state machines to control ankle mechanics during walking. State-based controllers are suitable for stereotypical cyclic locomotor tasks (e.g. walking on level ground) where joint mechanics are well defined at specific gait phases (i.e. states) and state transitions are easily detected. However, state-based controllers are not ideal for non-stereotypical acyclic tasks (e.g. freestyle dancing) where joint mechanics cannot be predefined and transitions are unpredictable. An alternative to state-based control is to utilize the amputee's nervous system for myoelectric control. A robotic lower limb prosthesis that uses continuous proportional myoelectric control would allow the amputee to adapt their ankle mechanics freely. One potential source for myoelectric control is the amputee’s residual muscles. I conducted four studies to examine the feasibility of using residual muscles for continuous myoelectric control during walking.In my first study, I demonstrated that it is possible to record residual electromyography from amputees during walking that are viable for continuous myoelectric control. My results showed that the stride-to-stride variability of residual and intact muscle activation patterns was similar. However, residual muscle activation patterns were significantly different across amputee subjects and significantly different than corresponding muscles in intact subjects. In my second study, I built and tested an experimental powered transtibial prosthesis and demonstrated that an amputee subject was able to walk using continuous proportional myoelectric control to alter prosthetic ankle mechanics. In my third study, I showed that five amputee subjects were able to adapt their residual muscles to walk using continuous proportional myoelectric control. With visual feedback of their control signal, amputees were able to generate higher peak ankle power walking with the experimental powered prosthesis compared to their prescribed prosthesis. In my fourth study, I conducted a user experience study and found that despite challenges with the device user interface, walking with continuous proportional myoelectric control gave amputees a sense of empowerment and embodiment. The results of my studies demonstrated the advantages and disadvantages of using continuous proportional myoelectric control for a powered transtibial prosthesis and suggest how next generation prostheses can build upon these findings.