DefinePK

Abstract

Over the last decades, the robotics industry has evolved exponentially, and humanoid robots cannot only be made but also, they can perform the physical functions of people. From this point of view, robotic hands are vital for many people who suffer either from an amputation or from any other disease. The main objective of this research was to classify the Electromyography (EMG) signals received from the human arm of healthy people and then carry out the manual application with a robotic hand in a virtual environment. This is especially important to understand and classify the geometric structure of the object contained in robotic handheld applications. The classification time and the precision relationship between the Artificial Neural Networks (ANN) and the Extreme Learning Machines (ELM) used for this classification were investigated. For this, 10 characteristics were extracted and the classifications were tested using ANN and ELM. The successful classification results obtained were compared with each other and applied to a virtual robotic hand using the V-Rep program.