Pengenalan Suara Untuk Penggerak Robot Lengan Dengan Metode LPC Dan ANN Menggunakan Perangkat Raspberry Pi 3

Abstract

At this time, technological developments have progressed so rapidly, one of which is in robotics with speech recognition as a robot mover. The use of speech recognition for interactively controlled robotic propulsion can replace the robot control system which is still manual. For this reason, the design and manufacture of a robotic arm movement system with the introduction of voice as a control. This study designed a system so that the robotic arm can be moved based on the voice commands intended for use handicapped (no hands) that utilizes voice signal extraction techniques by taking the pattern of voice commands generated by LPC and ANN methods. Based on the design results, controlling the robotic arm with voice recognition can be well detected. LPC method is used as feature extraction for speech recognition, while ANN with backpropagation algorithm is used for decision making from the resulting sound pattern. This study was tested with ten samples for each of the four speech recognition given. The study results show that the system can recognize the given sounds with an accuracy of 90% with a total of 40 test words from 4 types of words, namely right, left, clamp and release.
Keyword: Voice Recognition, Arm Robot, Linear Predictive Coding, Artificial Neural Network.