AI-Powered IoT Anomaly Detection

Bhangire, Amey Manishkumar (2025) AI-Powered IoT Anomaly Detection. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (842kB) | Preview
Preview	PDF (Configuration Manual) Download (1MB) | Preview

Abstract

This project focuses on detecting unusual behavior in IoT sensor data using a mix of simple and smart techniques. With more devices collecting data, it is important to make sure that the data is reliable and not affected by errors or attacks. In this study, we used the Intel Berkeley Lab dataset and added fake errors to test our models. We applied three main methods — Isolation Forest, One-Class SVM, and an Autoencoder—to find these unusual patterns. We also used a reinforcement learning agent to help make better decisions over time. All models were tested using standard accuracy checks like precision, recall, and F1-score. One-Class SVM performed the best, especially in finding real problems. We built the system to run in real time using a simple Streamlit interface, and it can also work on small edge devices like Raspberry Pi. This project offers a smart and useful solution to keep IoT systems safe by finding problems early and taking the right steps quickly.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Thomas, Lavish UNSPECIFIED
Subjects:	Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence T Technology > TK Electrical engineering. Electronics. Nuclear engineering > Telecommunications > Computer networks > Internet of things Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Artificial Intelligence
Depositing User:	Ciara O'Brien
Date Deposited:	11 Aug 2026 15:31
Last Modified:	11 Aug 2026 15:31
URI:	https://norma.ncirl.ie/id/eprint/9498

Actions (login required)

View Item