Publication Date: 2024/10/10
Abstract: This paper presents an innovative approach for at- tack classification on Unmanned Aerial Vehicles (UAVs) using the ALBERT (A Lite BERT) transformer model. As UAVs become in- tegral to various applications, their vulnerability to cyberattacks poses significant security challenges. Traditional methods often struggle with detecting sophisticated and evolving threats. By leveraging ALBERT’s efficiency in handling large-scale data, this study enhances the detection and classification of various UAV attack types. We describe the system model, problem formulation, and the proposed ALBERT- based classification framework. The model’s performance is evaluated through experimental results, demonstrating improvements in accuracy, precision, and recall compared to existing methods. The findings underscore the po- tential of transformer-based models in cybersecurity, specifically in safeguarding UAV systems. This work also opens avenues for future research into broader applications of ALBERT in other cybersecurity domains. The proposed framework offers a practical solution for enhancing UAV security in real-world scenarios.
Keywords: UAV, Attack Classification, ALBERT Transformer, Deep Learning, Cybersecurity.
DOI: https://doi.org/10.38124/ijisrt/IJISRT24SEP791
PDF: https://ijirst.demo4.arinfotech.co/assets/upload/files/IJISRT24SEP791.pdf
REFERENCES