A Learning-Based Fault Detection Approach for Unmanned Aerial Systems Using Inertial Sensor Data
8th International Congress on Human-Computer Interaction, Optimization and Robotic Applications, ICHORA 2026, Ankara, Turkey, 21 - 23 May 2026, (Full Text)
- Publication Type: Conference Paper / Full Text
- Doi Number: 10.1109/ichora69329.2026.11537197
- City: Ankara
- Country: Turkey
- Keywords: Embedded Systems, Fault Detection, Real-time Systems, Unmanned Aerial Vehicles, Unsupervised Learning
- Hacettepe University Affiliated: Yes
Abstract
Current fault detection methods often exceed the computational limits of small-scale unmanned systems, creating a need for lightweight, sensor-efficient diagnostic solutions. In this paper, we introduce a low-cost, embedded fault detection approach for unmanned systems based on inertial measurement unit (IMU) data. We focus on a quadrotor unmanned aerial vehicle (UAV) using a micro-electromechanical systems (MEMS) IMU. Our method employs an unsupervised, anomaly-based learning algorithm to model healthy operational behavior from sensor outputs. The system identifies deviations from this behavior as potential faults. Designed for resource-limited platforms, the algorithm operates in real-time on an onboard microcontroller. Experimental results demonstrate that the system achieves reliable diagnostics using only IMU data, eliminating the need for labeled fault samples or complex sensor fusion.