Investigation of Frequency-Domain-Based Vibration Signal Analysis for UAV Unbalance Fault Classification
Luttfi A. Al-Haddad, Alaa Abdulhady Jaber, Paramin Neranon, Sinan A. Al-Haddad
Abstract
The flying reliability of Unmanned Aerial Vehicles (UAVs) and flying robots, which directly determines the operational degree of safety, is becoming more important in recent intelligent decades. Reliability and a high level of safety are critical for autonomously controlled flying robots, especially in transportation and entertainment applications. Subsequently, monitoring UAV health is crucial and essential for system safety, cost savings, and excellent dependability. The development of numerous monitoring strategies has resulted from the requirement for a simple and accurate unbalance classification procedure. This paper provides an Unbalance Classification and Isolation (UCI) system for multirotor UAV propeller impairments. The technique is based on the processing of signal vectors from a vibration sensor positioned in the lines of the intersection of a modern-day drone's four propulsion units, which supply data for the Fast Fourier Transform (FFT) feature extraction. To identify and locate broken blades, characteristic fault signatures collected from vibration signals are employed and displayed in real-time on the programming platform. A noticeable maximum frequency shifting percentage value of 4.2% is acquired when deviating from a healthy state. The results reveal that identifying and isolating defective rotor states has high sensitivity and outperforms current studies in regard to unbalance classification of UAVs. The adopted technique is an efficient and low-cost solution that can be implemented in any multirotor UAV.