ABSTRACT
Electromyography, the recording of muscular activity, is of importance in industrial, biomechanical, and sports research as well as in medical diagnoses. When a muscle is activated, an electric potential in the order of microvolts (muV) is generated. This potential can be picked up, amplified, and displayed on an oscilloscope or strip chart recorder. Researchers have developed ways of analyzing these signals in terms of their characteristics. A numerical index, which reflects the basic characteristics of the electromyogram, mainly amplitude, frequency, and duration, can be used to provide quantitative information. The method used in this work for EMG processing consisted of filtering, rectification, and integration over very small intervals of time. Both analog and digital filtering proved necessary. Angular accelerometer and rotational potentiometer data were used in conjunction with limb inertia parameters obtained from existing biochemical models for the individual tested for obtaining the torques as a function of time. A system parameter identification method was used to determine the muscle parameters, such as occur in the single muscle Hill model, of four muscle groups for a human arm. The main results consist of personalized arm muscle group models. It was concluded that the method provided excellent (fit) personalized arm muscle group models under dynamic conditions. This method could lead to fundamental scientific information about a living muscle group from experiments in vivo.
