Kendali Mobile Robot Berbasis Pose Tangan Menggunakan Convolutional Neural Network (CNN)

Arzaqi, Alfan Miftah (2023) Kendali Mobile Robot Berbasis Pose Tangan Menggunakan Convolutional Neural Network (CNN). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 07211940000003-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 September 2025. Download (19MB) | Request a copy

Abstract

Awalnya robot hanya dapat dikendalikan secara dekat, namun beberapa tahun berikutnya robot sudah bisa dikendalikan dengan jarak yang jauh dengan tanpa kabel atau wireles dan dikendalikan dengan remote control. Kendali robot menggunakan pose tangan mulai dikembangkan menggunakan sensor yang diletakkan pada tangan dan juga dapat menggunakan computer vision untuk mengetahui pose tubuh manusia. Pengembangan kendali robot menggunakan pose tangan bertujuan untuk menciptakan kendali yang interaktif. Algoritma computer vision ada suatu metode yaitu Convolutional Neural Network (CNN) yang dapat memproses suatu data berupa citra dengan efisien. Tantangan pembuatan pose tangan untuk dikembangkan sebagai kendali robot yaitu terdapat variasi terhadap ukuran tangan manusia serta terdapat gangguan saat pengambilan citra tangan menggunakan kamera. Terdapat metode untuk mengenali pose tangan menggunakan landmark. Salah satu metode pembuatan landmark tangan adalah menggunakan framework medipipe. Maka da i itu dibuat kendali mobile robot berbasis pose tangan menggunakan Convolutional Neural Network (CNN). Nilai akurasi yang didapatkan dengan dari pengujian 500 data adalah 100% dan didapatkan jarak ideal antara 40cm sampai 120cm.
====================================================================================================================================
Originally, robots could only be controlled up close, but in the following years, they have become capable of being controlled remotely from a distance without cables or wires, using wireless technology and remote controls. The development of robot control using hand poses has started, employing sensors placed on the hand and computer vision to recognize human body poses. The aim of developing hand pose control for robots is to create an interactive control system. The computer vision algorithm, known as Convolutional Neural Network (CNN), efficiently processes image data. Challenges in creating hand poses for robot control include variations in human hand sizes and disturbances during image capture using cameras. There are methods available to recognize hand poses using landmarks, and one such method involves utilizing the Mediapipe framework. Thus, a mobile robot control system based on hand poses is developed using Convolutional Neural Network (CNN). The accuracy achieved from testing 500 data points is 100%, and an ideal range is obtained between 40cm to 120cm. The ideal condition occurs when the hand faces the camera.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Convolutional Neural Network, Mobile Robot, Pose
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) T Technology > TJ Mechanical engineering and machinery > TJ211.4 Robot motion T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5105.546 Computer algorithms
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Computer Engineering > 90243-(S1) Undergraduate Thesis
Depositing User:	Alfan Miftah Arzaqi
Date Deposited:	03 Aug 2023 15:37
Last Modified:	29 Aug 2023 02:45
URI:	http://repository.its.ac.id/id/eprint/101444

Actions (login required)

View Item