Bioelectric Control of Prostheses

Externally powered prostheses have been studied for many years, in efforts to provide more effective rehabilitation of amputees. In order to take advantage of the benefits offered by external poser in prostheses, however, the mode of control of the prosthesis by tWe amputee must also be improved. Conventional control methods require a high degree of mental concentration by the amputee on his prosthesis, because of at least two important factors: (i) the mode of control of a prosthesis motion conventionally is different from control of the corresponling norms: action; and (Ii) there is no feedback of scnsation from the prosthesis to the patient except through the visual sense. Bioelectric conti ol offers the possibility that control of the prosthesis action can be similar to control of the cor-esponding body action, which was lost throuigh the amputation. The muscles in the body prod%, an electrical signal, cMlled the electromyogram (emg), which can be sensed directly from the -,rface of th-! skin. Several experimental prostheses have been developed previously, which use emg signals for control purposes. The major shortcoming of most of th, ,e .de',icehab been the lack of graded control of their behavior. A systtm has been developed which utilizes surface emg signals from the biceps and tricens of an rinputee's erm to provide graded control of an elbow prosthesis. The development included as an intermediate step the control of a simulated forearm in a digital computer, in real time. hi its present form the signal processing consists of full-wave rectification and lowpass filtering of the emg signals from Alceps and triceps muscles; an actual mechanical elbow prosthesis can now be voluntarily controlled through the subject's emg signals. The performance of the system in its present form is described, and its indications for future work are outlined. Foremost among these is the need for feedback to the patient of (at least) position information from the prosthesis, outside of the visual se, ae. One method by which it may be possible to accomplish this in an inherently normal manner is suggested. The use of nerve signals as the control signal for a prosthes.•s offers many potential advantages over the use of the emg signal; the practical problems involved in observing nerve signals, however, combined with the lack of information as to how to properly interpret them, makes this approach infeasible now. Preliminary experiments aimed at deriving the necessary information in order that nerve signals can ultimately be used are described here. The results are still inconclusive.