Sistem Pengendali Robot Beroda Menggunakan Perintah Suara Manusia Berbasis Teknologi Tinyml

Abstract

Controlling the direction of the robot with voice commands to carry out an automatic recognition process on what is ordered by someone based on recognition in the form of a sound signal is proposed in this study to create a practical and efficient method so that the robot can move according to the user's instructions to order the robot to move without having to touch the robots. TinyML is a concept or term that refers to the use of a Machine Learning (ML) model that is very small and efficient in size and computational power, so that it can be implemented on a computer. In the voice signal processing control system by recording sound using the Arduino Nano 33 BLE Sense tool, then the data is connected to the Edge Impulse application to perform the recording. In the robot there is a microcontroller to give orders so that it is expected to be able to move the direction of the wheeled robot. The Convolutional Neural Network method is used so that the system can recognize each voice command delivered based on the features obtained. In this study, 5 voice command words are used, namely Forward, Backward, Right, Left, and Stop. Tested 5 times, the robot will move according to the command given with a success rate of 78.28%.