Perbandingan Neural Network Backpropagation dan Extreme Learning Machine pada Robot Manipulator

Abstract

Robotics technology and artificial intelligence devices are growing rapidly in the medical field. The work of medical workers will be made easier by the presence of this technology. It is also applied directly to patients. Patients suffering from various diseases, of course, need an appropriate solution. A robotic finger manipulator can be applied to a patient with disabilities to assist him in placing or retrieving items. In the manipulator robot, there are sensors, controllers, and actuators. The stimulation of the muscles in the forearm is detected by an electromyograph sensor. The resulting muscle stimulation is classified by the controller into servo motor movement. The motor represents the fingers. The classification method uses a neural network backpropagation and an extreme learning machine which is compared to the performance. Classification using neural network backpropagation has a success rate of 65.3%. While the classification using the extreme learning machine has a success rate of 78.7%.

Keywords

• artificial intelligence, robot manipulator, electromyograph, motor servo, neural network backpropagation, extreme learning machine