Establishing a Simplified Functional Relationship between EMG Signals and Actuation Signals Using Artificial Neural Networks

Using EMG signals as control signals has been a widely accepted option in the last decades. Using a wide array of techniques, EMG signals can be used in a variety of practical ways, from prostethics to exoesqueletons, however a concrete functional relationship between EMG signals and the dynamic and kinematic aspects of the upper limbs has not been established. Nowadays, almost every device that uses EMG signals uses them for classification purposes. In this work, we employ Fourier analysis in conjunction with other signal processing tools to treat the EMG signal, the treated signal is then used as an input of an artificial neural network in order to establish a simplified functional relationship between EMG and the upper limbs. We also employed other traditional signal processing methods for comparison purposes.