Comparison of ELM and ANN on EMG Signals Obtained for Control of Robotic-Hand

Abstract

The intelligent robotics industry is evolving and humanoid robots can be made, as well as being able to perform the physical functions of people. Robotic hands are vital for many people at this point. In this study, it was aimed to classify the Electromyography (EMG) signals received from the human arm in the correct direction and then to perform robotic hand application. This is very important in understanding and classifying the geometric structure of the object held in robotic hand applications. The classification time and accuracy ratio between ANN and ELM used for this classification were investigated. For this, 11 features were extracted and classifications were tried by using Extreme Learning Machine (ELM) and Artificial Neural Networks (ANN). The obtained successful classification results were compared with each other and applied to a robotic-hand.