RANCANG BANGUN APLIKASI ANDROID DETEKSI SAMPAH LOGISTIK PENDAKI GUNUNG DENGAN PENDEKATAN OFFLINE-FIRST, CIRCUIT BREAKER, DAN CRDTs

Kammala, Haffat Ahilla (2025) RANCANG BANGUN APLIKASI ANDROID DETEKSI SAMPAH LOGISTIK PENDAKI GUNUNG DENGAN PENDEKATAN OFFLINE-FIRST, CIRCUIT BREAKER, DAN CRDTs. Undergraduate thesis, UPN Veteran Jawa Timur.

	Text (Cover) 21081010234-Cover.pdf Download (1MB)
	Text (Bab 1) 21081010234-Bab1.pdf Download (155kB)
	Text (Bab 2) 21081010234-Bab2.pdf Restricted to Repository staff only until 27 May 2027. Download (1MB)
	Text (Bab 3) 21081010234-Bab3.pdf Restricted to Repository staff only until 27 May 2027. Download (4MB)
	Text (Bab 4) 21081010234-Bab4.pdf Restricted to Repository staff only until 27 May 2027. Download (12MB)
	Text (Bab 5) 21081010234-Bab5.pdf Download (139kB)
	Text (Daftar Pustaka) 21081010234-Daftar Pustaka.pdf Download (121kB)

Abstract

Mountain climbing is a popular activity enjoyed by many people; however, it often leaves behind environmental issues, particularly in the form of improperly discarded logistical waste. This research focuses on designing and developing an Android application to detect mountaineering logistical waste using an Offline-First approach, the Circuit Breaker pattern, and Conflict-free Replicated Data Types (CRDTs). The Offline-First approach enables the application to function without an internet connection, the Circuit Breaker ensures system stability during repeated communication failures, and CRDTs allow consistent data synchronization across devices. The application integrates an object detection model based on YOLO, trained with datasets processed via Roboflow. The backend is developed using Go and PostgreSQL, while SQLite is used for local storage. The implementation results show that the system successfully records mountaineering logistical waste using the Offline-First design, allowing the Android application to operate optimally even without internet access through the use of SQLite and asynchronous data synchronization. Object detection is performed using the YOLO model integrated via the Roboflow API. The application employs the BLoC architecture to separate the user interface from business logic and utilizes a DatabaseHelper for offline CRUD operations. The implementation of the Circuit Breaker pattern improves system reliability by handling network failures through three primary connection states. Additionally, the use of CRDTs with a vector clock mechanism ensures data consistency across devices, even during concurrent updates in unstable network conditions.

Item Type:

Thesis (Undergraduate)

Contributors: