Image of IMPLEMENTASI DETEKSI OBJEK PADA HUMANOID ROBOT BERBASIS ALGORITMA YOU ONLY LOOK ONCE (YOLO)

Skripsi

IMPLEMENTASI DETEKSI OBJEK PADA HUMANOID ROBOT BERBASIS ALGORITMA YOU ONLY LOOK ONCE (YOLO)

Penilaian

4

dari 5
Penilaian anda saat ini :  

Nowadays, the research in robots has grown fast. Even robots can have constructions that resemble humans, including interacting with the surrounding environment. The robot can perform this interaction through a camera that can process visual input to recognize objects around it. However, the current object recognition method has not been implemented in real-time on humanoid robots. Thus, this study developed a system that can detect various objects surrounding a humanoid robot using the You Only Look Once (YOLO) algorithm. In this study, the objects detected by the humanoid robot consisted of 12 classes, namely humans, chairs, tables, cabinets, computers, books, doors, bottles, eggs, learning modules, cups, and hands. The object dataset used in this study is primary data taken at the Robotics and Control Engineering laboratory. Each object class consist of 1500 data. 2 types of YOLO architecture were used in this research, namely Tiny YOLOv3 and Tiny YOLOv4. This aims to see the performance of the two architectures in detecting objects around the robot. The results showed that the Tiny YOLOv4 model had better performance than the Tiny YOLOv3. The detection accuracy in the simulation experiment obtained was 74,16%. Then, the accuracy in real-time conditions was 61,66% for bright and 38,33% for dim condition. It showed that Tiny YOLOv4 has a good performance on object detection systems around humanoid robots, especially for bright condition. The system built also sent data information in the form of class, distance, and object position used by robots to interact with humans.

Availability
Inventory Code Barcode Call Number Location Status
2207003276T72888T728882022Central Library (Referens)Available
Detail Information
Series Title
-
Call Number
T728882022
Publisher
Indralaya : Prodi Teknik Elektro, Fakultas Teknik Universitas Sriwijaya.,
Collation
xiii, 64 hlm.; ilus.; 29 cm
Language
Indonesia
ISBN/ISSN
-
Classification
629.807
Content Type
-
Media Type
-
Carrier Type
-
Edition
-
Subject(s)
Teknik elektro
Teknik Kontrol Otomatis, robotika
Specific Detail Info
-
Statement of Responsibility
Other version/related

No other version available

File Attachment
Comments
You must be logged in to post a comment