Abstract
In human-prosthesis systems, the signals must be transmitted through the functions remained in the amputee. Therefore, the amputee's ability of information transmission is largely restricted and it seems that the control ability may be inferior to the one in regular man-machine systems. It is important to identify human operator's control strategy and to clarify the limit of control ability under the restricted conditions for the development of powered prosthesis.
The purpose of this paper is to analyze human operator's control characteristics in man-machine systems controlled by quantized EMG signal. Human operator was asked to compensate a random input by manipulating quantized EMG signal with five states. First, human operator's describing functions were estimated from the experiments and it was shown that the gain characteristics became worse than the ones in regular man-machine systems. Then, human operator's switching behavior was analyzed and it was represented that the nonlinear property of the human operator led to the reduction of the gain characteristics.
