Advanced Fault Detection and Autonomous Drone SaFeTy System (ADAFT)
Recently, unmanned aerial vehicles (UAVs) or drones have played an important role in many industrial tasks, especially in tasks that are out of human reach or dangerous for humans, such as object transportation, routine inspection, search-and-rescue, and site/building exploration. Among these UAVs, quadrotors have gained significant attention due to their maneuverability, versatility, and ease of control. However, as with any complex system, quadrotors are susceptible to faults and failures that can affect their operation and pose risks to both property and lives.
To address these challenges and prevent drones from falling, one way is to develop a diagnostic monitoring system and control strategy platform to detect, estimate, and react to fault events. Advanced fault detection (FD) and fault-tolerant control (FTC) techniques have been proposed as main strategies for enhancing the safety and reliability of quadrotor systems. By integrating sophisticated algorithms and sensor systems, these approaches enable real-time monitoring of the quadrotor’s health status, timely detection of faults, and seamless adaptation to mitigate their effects.
Following these strategies, this project focuses on developing a state-of-the-art fault detection algorithm using data from multiple sensors to detect and diagnose various types of anomalies (e.g., sensor failures, actuator malfunctions, or mechanical damage). Leveraging insights from control theory, machine learning, and aerospace engineering, the aim is to also create robust fault-tolerant strategies specifically tailored for quadrotors to ensure their stability, safety (safe landing), and performance under adverse conditions. These two aspects will lead to Advanced Fault Detection and Autonomous Drone Safety Systems (ADAFT).
Official website: https://adaft.vistec.ac.th/
