Unmanned Aerial Vehicles (UAVs), commonly referred to as drones, have determined numerous applications across industries, ranging from aerial surveillance to package shipping. As drones are used in vital operations, ensuring cyberattacks and anomalies protect them from attackers is now a big challenge. This research study presents a complete approach to enhancing drone safety by integrating multi-sensor anomaly detection and superior machine learning techniques. The proposed methodology capitalizes on the rich sensor suite embedded in present-day drones, encompassing GPS receivers, accelerometers, gyroscopes, cameras, communication modules, and more. Leveraging an array of sensors in drones, our technique detects abnormal drone behavior indicative of unauthorized access, GPS spoofing, communication jamming, and malicious activities. By extracting features from sensor records, we develop a robust anomaly detection framework using the “uav attack dataset” able to identify deviations from normal flight patterns, communication signals, and environmental interactions. Central to our methodology is the utilization of machine learning algorithms. These algorithms are skilled on labeled datasets containing numerous flight eventualities, each normal and hostile, together with the ones discovered inside the “uav attack dataset”. The obtained results are eventually evaluated using rigorous performance metrics to quantify their effectiveness in distinguishing genuine anomalies from benign variations. The findings of our study underscore the capacity of multi-sensor anomaly detection for drones. By harnessing the power of machine learning and sensor fusion, we exhibit the ability to hit upon attacks at an early level, mitigating capability harm and permitting rapid responses. This study contributes now not only to the field of drone safety but also to the broader panorama of self-sustaining systems protection, highlighting the importance of adaptive and proactive protection mechanisms. Results show that an accuracy of 99% with AUC of 100$ was achieved when all the sensors were used.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Enhancing Drone Security Through Multi-Sensor Anomaly Detection and Machine Learning


    Additional title:

    SN COMPUT. SCI.


    Contributors:

    Published in:

    Publication date :

    2024-06-13




    Type of media :

    Article (Journal)


    Type of material :

    Electronic Resource


    Language :

    English




    Anomaly detection via drone

    HILLER NATHAN D / NEWMAN DAVID I / TORRES ROBERT B | European Patent Office | 2024

    Free access

    ANOMALY DETECTION VIA UNMANNED AERIAL DRONE

    HILLER NATHAN D / NEWMAN DANIEL I / TORREZ RAYMUNDO B | European Patent Office | 2024

    Free access

    ANOMALY DETECTION VIA UNMANNED AERIAL DRONE

    HILLER NATHAN D / NEWMAN DANIEL I / TORREZ RAYMUNDO B | European Patent Office | 2024

    Free access