Prayoga, Moh. Harwin (2021) Rancang Bangun Sistem Deteksi Rintangan Pada Prototipe Autonomous Car Menggunakan Sensor Fusion Berbasis Kamera Mono Dan Sensor Ultrasonik. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 02311740000122-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
02311740000122-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2023. Download (2MB) | Request a copy |
|
Text
02311740000122-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (2MB) | Request a copy |
Abstract
Rancang bangun sistem deteksi ini dirancang dengan
algoritma sensor fusion untuk mendeteksi jarak, koordinat, serta ukuran tinggi, dan lebar bounding box objek menggunakan kamera mono dan sensor jarak ultrasonik. Kamera mono berfungsi untuk menyediakan informasi pixel gambar dan sensor ultrasonik berfungsi untuk menyediakan informasi jarak. Kamera yang digunakan adalah kamera mono yang memiliki field of view (FOV) sebesar 160º. Metode deteksi yang digunakan pada kamera mono tersebut adalah YOLOv4-Tiny. Sensor jarak yang digunakan adalah sensor ultrasonik yang memiliki FOV sebesar 15º. Performansi dari metode deteksi YOLOv4-Tiny menghasilkan delay sebesar 7 detik sehingga perlu dilakukan output delay pada sensor ultrasonik agar dapat menghasilkan data yang sinkron dengan deteksi objek pada kamera. Pengujian dilakukan dengan menggunakan dua kasus yang berbeda yaitu objek tunggal dengan keadaan statis dan dinamis. Hasil dari validasi masing-masing sensor pada jarak 20 cm hingga 60 cm menunjukkan kemampuan deteksi dengan error rata-rata kurang dari 5%. Sistem deteksi mampu mendeteksi objek dengan benar pada jarak hingga 60 cm. Algoritma sensor fusion yang dibangun mampu menghasilkan informasi tinggi dan lebar bounding box objek dengan nilai ketidakpastian berturut-turut 0,05 cm dan 0,06 cm; informasi jarak dengan ketidakpastian 0,76 cm; serta informasi posisi lateral dengan ketidakpastian 0,03 cm
=====================================================================================================
The design of this detection system is designed with a fusion sensor algorithm to detect the distance, coordinates, as well as the height, and width bounding box of objects using mono cameras and ultrasonic distance sensors. Mono cameras serve to provide image pixel information and ultrasonic sensors serve to provide distance information. The camera used is a mono camera that has a field of view (FOV) of 160º. The detection method used in the mono camera is YOLOv4-Tiny. The distance sensor used is an ultrasonic sensor that has an FOV of 15º. Performance of the YOLOv4-Tiny detection method produces a delay of 7 seconds so it is necessary to perform output delay on ultrasonic sensors in order to produce data that is in sync with the detection of objects on the camera. The test was conducted using two different cases: a single object with a static and dynamic state. Results from the validation of each sensor at a distance of 20 cm to 60 cm showed detection capability with an average error of less than 5%. The detection system is able to detect objects correctly at distances of up to 60 cm. The fusion sensor algorithm is capable of producing information on the height and width of bounding boxes of objects with uncertainty values of 0.05 cm and 0.06 cm, consecutively. distance information with uncertainty of 0.76 cm; as well as lateral position information with uncertainty of 0.03 cm
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Uncontrolled Keywords: | Sensor Fusion, Kamera Mono, Sensor Ultrasonik, YOLOv4-Tiny |
| Subjects: | T Technology > TL Motor vehicles. Aeronautics. Astronautics > TL152.8 Vehicles, Remotely piloted. Autonomous vehicles. |
| Divisions: | Faculty of Industrial Technology and Systems Engineering (INDSYS) > Physics Engineering > 30201-(S1) Undergraduate Thesis |
| Depositing User: | Moh. Harwin Prayoga |
| Date Deposited: | 03 Sep 2021 00:57 |
| Last Modified: | 03 Sep 2021 00:57 |
| URI: | http://repository.its.ac.id/id/eprint/91399 |
Actions (login required)
 |
View Item |