Article: Enhancing autonomous vehicle security in intelligent transportation systems through quantum computing and optimisation based federated learning Journal: International Journal of Heavy Vehicle Systems (IJHVS) 2025 Vol.32 No.6 pp.776 - 797 Abstract: Autonomous vehicles (AVs) and smart transportation systems have the ability to change modern society, but confidentiality and security must be guaranteed. The ecosystems of AVs employ machine learning models, which are susceptible to adversarial attacks during training by federated learning and data poisoning. Hyperparameters are vital for creating a resilient federated learning model. The proposed system addresses these issues by employing adversarial attacks to automatically adjust hyperparameters. The methodology consists of two phases: updating learning rate hyperparameters during global and local epochs, and designing a framework to defend against attacks. The proposed system is evaluated using two benchmark datasets, Fashion- Modified International Standards for Telecommunication (MNIST) and MNIST, enhancing AV security through real-time hyperparameter optimisation using quantum computing. This proposed system improves resistance to external attacks while safeguarding data privacy. Inderscience Publishers - linking academia, business and industry through research

Title: Enhancing autonomous vehicle security in intelligent transportation systems through quantum computing and optimisation based federated learning

Authors: S. Sellakumar; Kavin FrancisXavier; Pradeepa Karuppaiah; N. Bharathiraja

Addresses: Department of Mechanical Engineering, Saveetha Engineering College, Channai, 602105, Tamil Nadu, India ' Instrumentation Engineer, M/S Muscat Engineering Consultancy Pvt. Ltd., Trichy, Tamil Nadu, 620001, India ' Department of Computer Science and Engineering, Mahalakshmi Tech Campus, Chromepet, Chennai 600044, Tamil Nadu, India ' Department of Computer Science and Engineering, Dayananda Sagar University, School of Engineering, Bangalore 562112, Karnataka, India

Abstract: Autonomous vehicles (AVs) and smart transportation systems have the ability to change modern society, but confidentiality and security must be guaranteed. The ecosystems of AVs employ machine learning models, which are susceptible to adversarial attacks during training by federated learning and data poisoning. Hyperparameters are vital for creating a resilient federated learning model. The proposed system addresses these issues by employing adversarial attacks to automatically adjust hyperparameters. The methodology consists of two phases: updating learning rate hyperparameters during global and local epochs, and designing a framework to defend against attacks. The proposed system is evaluated using two benchmark datasets, Fashion- Modified International Standards for Telecommunication (MNIST) and MNIST, enhancing AV security through real-time hyperparameter optimisation using quantum computing. This proposed system improves resistance to external attacks while safeguarding data privacy.

Keywords: autonomous vehicles; intelligent transportation systems; quantum computing; federated learning; adversarial attacks; data privacy.

DOI: 10.1504/IJHVS.2025.150217

International Journal of Heavy Vehicle Systems, 2025 Vol.32 No.6, pp.776 - 797

Received: 06 Nov 2024
Accepted: 23 Feb 2025
Published online: 03 Dec 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Enhancing autonomous vehicle security in intelligent transportation systems through quantum computing and optimisation based federated learning

Keep up-to-date