Resilience of Sensor Fault Detection Systems Using LSTM for Intelligent Vehicle Systems

Sia, Lydia Jia Hui (2025) Resilience of Sensor Fault Detection Systems Using LSTM for Intelligent Vehicle Systems. Final Year Project (Bachelor), Tunku Abdul Rahman University of Management and Technology.

Text Lydia Sia Jia Hui_Full Text.pdf Restricted to Registered users only Download (4MB)

Abstract

Intelligent vehicles (IVs) are increasingly reliant on complex systems to ensure safety, efficiency, and convenience. To guarantee user safety, fault detection (FD) algorithms are essential. This study proposes a data-driven approach using LSTM regression models for fault detection in IV systems. By simulating various faults in velocity data and evaluating the model's performance, we demonstrate its effectiveness in accurately detecting anomalies. The evaluation metrics, including R-squared, Mean Absolute Error, Mean Squared Error, and Root Mean Squared Error, confirm the model's high accuracy and minimal prediction errors. This research highlights the potential of LSTM regression for fault detection in dynamic systems, providing a foundation for future applications in predictive maintenance and anomaly detection.