Fauzan, Muhammad Alif (2023) Deteksi dan Pengenalan Rambu Lalu Lintas Menggunakan Algoritma Deep Learning Pada Mobil Otonom. Other thesis, Institut Teknologi Sepuluh Nopember.
Text
07111940000189-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (3MB) | Request a copy |
Abstract
Berdasarkan data WHO (World Heath Organization), setiap tahunnya sekitar 1.3 juta lebih orang meninggal dikarenakan kecelakaan lalu lintas. Maka dari itu diperlukan solusi untuk meningkatkan keselamatan dan kenyamanan saat berkendara dalam hal lalu lintas. Maka dari itu dibuat sistem deteksi rambu lalu lintas menggunakan algoritma YOLO pada mobil otonom untuk dapat mendeteksi rambu lalu lintas. Algoritma YOLO ini merupakan perkembangan dari algoritma Convolutional Neural Network (CNN) dimana YOLO dapat mendeteksi objek dengan kecepatan mendekati real time. Sistem ini juga dilengkapi oleh estimasi jarak dari mobil otonom ke rambu lalu lintas. Kemudian Sistem ini dapat melakukan aksi pengereman pada mobil otonom ketika ada rambu batas kecepatan yang dideteksi dan dikenali sehingga dapat mematuhi peraturan rambu lalu lintas. Dalam proses mendeteksi dan mengenali objek terkadang ada beberapa faktor yang memengaruhi performa model deep learning, salah satunya adalah blur dan adanya noise. Biasanya noise muncul ketika pencahayaan rendah sehingga mengganggu proses deteksi objek dikarenakan gambar yang menjadi kurang jelas. Oleh karena itu digunakan filter wiener untuk merestorasi citra sehingga model dapat mendeteksi objek dengan lebih jelas. Simulasi yang digunakan untuk aksi pengereman mobil pada model deep learning rambu batas kecepatan menggunakan Simulink untuk memodelkan dinamika mobil sederhana, serta menggunakan kontroller Proporsional pada PID untuk mengontrol perlambatan mobil sehingga dapat memenuhi aturan batas kecepatan pada rambu batas kecepatan.
================================================================================================================================
Every year around 1.3 million more people die due to traffic accidents. Therefore a solution is needed to improve safety and comfort when driving in terms of traffic. Therefore a traffic sign detection system was created using the YOLO algorithm for autonomous car to be able to detect traffic signs. The YOLO algorithm is a development of the CNN algorithm where YOLO can detect objects at near real time speeds. This system is also equipped with an estimation of the distance from the autonomous car to traffic signs. Then this system can perform braking actions on autonomous cars when there are speed limit signs that are detected and recognized so that they can comply with traffic sign regulations. In the process of detecting and recognizing objects, several factors affect the performance of deep learning models, one of which is blur and noise. Usually noise appears when the lighting is low so that it interferes with the object detection process because the image becomes less clear. Therefore a wiener filter is used to restore the image so that the model can detect objects more clearly. The simulation used for car braking action on the speed limit sign deep learning model uses Simulink to model simple car dynamics, and uses a Proportional controller on PID to control car deceleration so that it can comply with the speed limit rules on speed limit signs.
Item Type: | Thesis (Other) |
---|---|
Uncontrolled Keywords: | Convolutional Neural Network, Mobil Otonom, YOLO, Estimasi Jarak, Wiener Filter, PID Controller, Autonomous Vehicle, YOLO, Distance Estimation, Wiener Filter |
Subjects: | Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QA Mathematics > QA336 Artificial Intelligence T Technology > TL Motor vehicles. Aeronautics. Astronautics > TL152.8 Vehicles, Remotely piloted. Autonomous vehicles. |
Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20201-(S1) Undergraduate Thesis |
Depositing User: | Muhammad Alif Fauzan |
Date Deposited: | 25 Jul 2023 08:27 |
Last Modified: | 25 Jul 2023 08:27 |
URI: | http://repository.its.ac.id/id/eprint/99312 |
Actions (login required)
View Item |