Prototipe Sistem Deteksi Maling Menggunakan Wi-Fi Based Secure Indoor Positioning System

Ramadhan, Naufal (2024) Prototipe Sistem Deteksi Maling Menggunakan Wi-Fi Based Secure Indoor Positioning System. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5027201067_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2026. Download (4MB) | Request a copy

Abstract

Sistem pemosisian dalam ruangan sudah mulai berkembang dalam beberapa tahun belakangan ini. Berbeda dengan sistem pemosisian yang sudah sangat dikenal oleh masyarakat yakni Global Positioning System (GPS), pada penelitian kali ini akan berfokus melakukan pemosisian di dalam ruangan atau biasa disebut dengan Indoor Positioning System (IPS). Indoor Positiong System (IPS) pada era digitalisasi sangat berpengaruh mengingat seluruh perkembangan infrastruktur sangat maju serta dengan sektor digitalisasi. Pada perkembangan di banyak sektor ini pula maka akan menimbulkan banyak masalah baru seperti contohnya apabila terdapat gedung pencakar langit maka terdapat banyak halangan atau masalah untuk memastikan lokasi dari orang atau pun benda. Masalah tidak berhenti sampai disitu, dimana ancaman juga dapat datang dari manusia baik masuk melalui sistem maupun secara fisik. Pada penelitian kali ini akan digunakan untuk mencari tau perbandingan model terbaik berdasarkan dataset yang sudah ada dengan dataset yang diambil secara langsung dalam salah satu ruangan yang dianggap steril. Kemudian dibandingkan menggunakan beberapa algoritma machine learning sebagai bahan perbandingan algoritma mana yang memiliki akurasi terbaik. Serta dibuat salah satu ujicoba metode serangan yang kemudian dibuat model yang robust. Kemudian dari perbandingan tersebut akan dipilih salah satu model dari dataset yang memiliki akurasi terbaik untuk dibuat prototipe sistem deteksi dimana apabila terdeteksi sebuah anomali di dalam model maka akan mengirimkan pesan warning melalui email dengan harapan sistem ini dapat berguna di ruang atau tempat restricted. Penelitian ini memanfaatkan dataset berbentuk Channel State Information (CSI). Dari beberapa model yang diuji diperoleh akurasi senilai 74,02% menggunakan algoritma random forest. Dimana implementasi serangan menggunakan Decision Tree Attack berhasil dilakukan hingga akurasi turun di angka 52%. Uji coba penerapan metode pertahanan menggunakan featuresqueezing berhasil membuat model lebig tahan dibuktikan dengan akurasi menjadi 68%. Serta implementasi Anomaly Detector dapat mengirimkan pesan bahaya melalui email.
=================================================================================================================================
Indoor positioning systems have begun to develop in recent years. Different from the positioning system that is well known to the public, namely the Global Positioning System (GPS), this research will focus on indoor positioning or what is usually called the Indoor Positioning System (IPS). The Indoor Positioning System (IPS) in the digitalization era is very influential considering that all infrastructure developments are very advanced as well as the digitalization sector. Development in many of these sectors will also give rise to many new problems, for example if there are skyscrapers then there will be many obstacles or problems in determining the location of people or objects. The problem doesn't stop there, where threats can also come from humans, either through the system or physically. In this research, it will be used to find out the comparison of the best model based on an existing dataset with a dataset taken directly in a room that is considered sterile. Then it is compared using several machine learning algorithms as a comparison which algorithm has the best accuracy. And a test method of attack was made which was then made into a sturdy model. Then, from this comparison, one of the models from the dataset that has the best accuracy will be selected to create a detection system prototype, where if an anomaly is detected in the model, it will send a warning message via email in the hope that this system can be useful in limited spaces or places. This research utilizes a dataset in the form of Channel State Information (CSI). Several models tested obtained an accuracy of 74.02% using the random forest algorithm. Where the implementation of the attack using Decision Tree Attack was successfully carried out until the accuracy fell to 52%. The trial application of the maintenance method using featurequeezing succeeded in making the model more durable, proven by an accuracy of 68%. And the Anomaly Detector implementation can send danger messages via email.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Algorithm, Indoor Positioning System, Machine Learning, Robust, Security, Algoritma, Keamanan
Subjects:	T Technology > T Technology (General) > T57.5 Data Processing T Technology > T Technology (General) > T57.8 Nonlinear programming. Support vector machine. Wavelets. Hidden Markov models.
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information Technology > 59201-(S1) Undergraduate Thesis
Depositing User:	Naufal Ramadhan
Date Deposited:	05 Feb 2024 17:25
Last Modified:	05 Feb 2024 17:25
URI:	http://repository.its.ac.id/id/eprint/106170

Actions (login required)

View Item