Status : Verified
| Personal Name | De Leon, Chloe S. |
|---|---|
| Resource Title | Privacy-Preserving Vehicle Intrusion Detection System Using Federated Learning and Homomorphic Encryption |
| Date Issued | 30 April 2026 |
| Abstract | As modern vehicles evolve into highly connected and autonomous systems, the need for robust cybersecurity measures, such as intrusion detection systems (IDS), becomes increasingly critical. Traditional IDS approaches, however, face significant challenges in preserving data privacy due to centralized collection and processing, which expose sensitive information to potential breaches. This study addresses these challenges by presenting a novel privacy-preserving vehicle IDS that combines federated learning (FL) and homomorphic encryption (HE). Specifically, the system detects denial-of-service (DoS), fuzzy, and impersonation attacks in vehicular networks to strengthen cybersecurity protections against common vehicular threats. FL enables collaborative model training across distributed vehicles without sharing raw data, ensuring data privacy. HE facilitates secure computations on encrypted data, enhancing confidentiality during processing. This study leverages the CKKS HE scheme to enable efficient encrypted aggregation of model updates, demonstrating the practical feasibility of deploying privacy-preserving collaborative learning solutions in real-world vehicular cybersecurity systems. However, the system is not intended for real-time detection. Due to the computational and communication overhead of FL and HE, this design is best suited for offline or near-real-time analysis where data privacy and collaborative diagnostics are prioritized. |
| Degree Course | Master of Science in Computer Science |
| Language | English |
| Keyword | vehicle intrusion detection system; federated learning; homomorphic encryption; cybersecurity; data privacy; connected vehicles |
| Material Type | Thesis/Dissertation |
Preliminary Pages
910.42 Kb
Category : P - Author wishes to publish the work personally.
Access Permission : Limited Access