Surface EMG signals pattern recognition utilizing an adaptive crosstalk suppression preprocessor

This paper proposes utilization of a Least Mean Square (LMS) based Finite Impulse Response (FIR) adaptive filter block, before conventional Surface Electromyogram (sEMG) signal pattern classification schemes. This novel configuration suppresses the sEMG between channels crosstalk. In this study, the sEMG signals are detected from the biceps and triceps brachii muscles to identify four primitive motions, i.e., elbow flexion/extension and forearm supination/pronation. A Multi Layer Perceptron (MLP) classifies the two time domain feature vectors that are extracted from raw and preprocessed sEMG signals, respectively. Although the implementation of an adaptive filter increases computational complexity, significant advances in sEMG pattern classification has been achieved.