Estimating the instantaneous wrist flexion angle from multi-channel surface EMG of forearm muscles

A pattern recognition based classification method is proposed to estimate wrist flexion angles from electrical activities of forearm muscles. Spatial movement data and multi-channel myoelectric signals from forearm muscles were collected experimentally during periodic wrist flexion and extension movements using an ultrasound based movement analyser system. The recorded marker coordinates were transformed into joint angles using OpenSim, an open source simulation tool for biomechanical analysis. EMG data were segmented according to specific ranges of the calculated wrist flexion angle to form different classes for pattern recognition. The parameter space of the used classification algorithm was explored with a selected subset of values to find the optimal parameter vector giving maximal classification performance.