Identification of Upper Limb Movements through EMG Signals with Fuzzy Logic Algorithm

Electromyography consists on recording the electrical potential generated by the activation of muscle fibers when performing voluntary or involuntary movements. Thus, EMG signals (EMGs) are directly linked to the human intention of motion. However, due to the random nature of EMGs; the correct prediction of the intention of motion is considered the most difficult part of the myoelectric control. A movement identification algorithm to distinguish among nine different movements of the upper limb is presented. Three fuzzy stages were developed. The if-then rules at each stage reflect the behavior of muscle fibers when performing a particular movement. This algorithm was evaluated using surface EMG recordings measured over the Deltoid, Bicep and Pronator Teres muscles. Three main features were extracted from each channel, the root mean square, the contractility characteristic and the onset value. Results have shown a high percentage of accuracy.