A clustering approach for attack detection and data transmission in vehicular ad-hoc networks

Network. 2023 Nov 18:1-26. doi: 10.1080/0954898X.2023.2279973. Online ahead of print.

Abstract

Vehicular ad-hoc networks (VANETs) are increasingly pivotal for empowering applications in smart cities and intelligent traffic systems. However, the reliability and stability of VANET communications face formidable obstacles. Natura 2000 (N2k), the largest globally coordinated network of protected areas, has drawn significant criticism for its conservation-centric management structure lacking a strategic vision. This study proposes a three-phase strategy to address these concerns, aiming for effective and sustainable N2K site management. The novel approach employs DNN-Assisted Canonical Correlation Analysis (DNN-CCAS), encompassing cluster formation, cluster head selection, and outbreak recognition for enhanced VANET security. Vehicle clustering begins with an amended K-consonance method, emphasizing both position and speed through AKCEM clustering. A cluster head is chosen via a linear measure promenade approach, followed by secure data transmission to the cloud using DNN-CCAS if the cluster head is deemed normal. The proposed method outperforms prevailing techniques, achieving an impressive 91% accuracy. This comprehensive strategy not only addresses VANET communication challenges but also aims to revolutionize the management of N2K sites by integrating a strategic vision into conservation practices.

Keywords: DNN-CCAS; Natura 2000; Vehicle ad hoc network; amended K-consonance expedient; linear measure promenade approach.