Extraction of elbow joint intention from sEMG signals in horizontal plane using cosine tuning functions

A human-robot interaction (HRI) system could be more effective if it is intelligent enough to accurately measure and predict humans´ moving intention and adapt their behaviors accordingly. By means of measuring biomechanical signals of human body to make such predication has been a challenging research topic. This paper presents results of our investigation to extract human´s elbow joint motion intention from surface electromyographic (sEMG) measurement. While a human subject´s limb is kept still on a horizontal table but subject to certain force applied to the hand, by changing the direction of the force applied, the biomechanical signals are simultaneously recorded from three sEMGs mounted on superficial elbow and shoulder muscles in real-time. Based on the measurement obtained, we study the relation between static sEMG signals and movement trends of the elbow in eight directions using cosine tuning functions. Genetic algorithm (GA) is applied to determine multivariate nonlinear regression coefficient to improve the prediction accuracy.