Anomaly Detection in Quadcopter Systems Using AI and Vibration Signal Processing

This study examines the implementation of artificial intelligence (AI) techniques for anomaly identification in quadcopters, concentrating on vibration data obtained from the ADXL345 accelerometer. The ESP32 Wi-Fi module analyses and transmits these signals, delivering high-resolution data appropriate for real-time monitoring. Diverse AI methodologies, encompassing supervised learning algorithms, are utilised to analyse the data and identify anomalies. Feature extraction techniques, including standard deviation (std), variance, and mean absolute deviation (MAD), are employed to augment the predictive capability of the data. Classifiers such as Random Forest and Support Vector Machine(SVM) are utilised on the retrieved features, with a maximum accuracy of 97.78%. The amalgamation of AI and IoT technologies facilitates an efficient and effective approach to enhance defect identification and problem diagnostics in unmanned aerial vehicles (UAVs), presenting substantial promise for augmenting the dependability and safety of these systems.