Anomaly Detection and Predictive Maintenance in IoT Gateways

Chan, Esmond Yu Hern (2026) Anomaly Detection and Predictive Maintenance in IoT Gateways. Final Year Project (Bachelor), Tunku Abdul Rahman University of Management and Technology.

Text ESMOND_CHAN_YU_HERN_Full Text.pdf Restricted to Registered users only Download (3MB)

Abstract

IoT gateways serve as intermediaries between edge devices and cloud platforms to ensure seamless data transmission. Traditional reactive approaches mean that anomalies like equipment faults and degradation can easily be missed and result in costly downtimes. Hence, real-time anomaly detection enables predictive maintenance to maintain reliability and ensure effective resource allocation. This study will focus on conducting a detailed literature review and comparative research on anomaly detection and predictive maintenance techniques in IoT gateways for monitoring equipment such as electrical motors, pumps and machines in an industrial setting. The research objectives aim to reviewexisting anomaly detection and prediction algorithms such as Random Forest, Logistic Regression, Support Vector Machine (SVM), LSTM and Autoencoders applied in industrial IoT settings to identify unusual patterns in real time. Furthermore, this research also aims to analyze these models to evaluate their effectiveness in predicting equipment failures. Finally, this project aims to contribute towards implementing efficient and real-time machine learning-based anomaly detection systems for IoT gateways. The research involved will provide an overall evaluation of existing methods and implementation feasibility, in order to guide future developments in efficient equipment monitoring and maintenance strategies