| Title | Author | Created | Published | Tags | | ------------- | ---------------------------- | -------------- | -------------- | -------------------------------------------------- | | Example Paper | <ul><li>Jon Marien</li></ul> | April 03, 2025 | April 03, 2025 | [[#classes\|#classes]], [[#SYST44998\|#SYST44998]] | # Securing In-Vehicle Communication Networks Against Cyber Threats **Jonathan Marien**† \*FAST Department -- Cyber Security, Sheridan College, Oakville, Canada [email protected] **Abstract**—This paper analyzes security vulnerabilities in modern in-vehicle networks using Controller Area Network (CAN) protocols and proposes a multi-layered defense framework. Through empirical testing of 2018-2025 vehicle models, we identify critical attack vectors in ECUs and propose a hybrid authentication mechanism combining lightweight cryptography with ML-based anomaly detection. **Index Terms**—Vehicle cybersecurity, CAN bus security, ECU authentication, Automotive intrusion detection ## I. Introduction Modern vehicles contain over 150 electronic control units (ECUs) communicating via legacy protocols like CAN [1]. Recent attacks demonstrate remote hijacking through: - Compromised infotainment systems [2] - OBD-II port exploitation [3] - TPMS sensor spoofing [4] Our contribution includes: 1. Threat model for CAN FD networks 2. Evaluation of post-quantum cryptographic solutions 3. Real-time anomaly detection framework ## II. Related Work | Approach | Strength | Limitation | |--------------------|-----------------------|--------------------------| | MAC-based auth [5] | Low latency | Vulnerable to replay | | ML detection [6] | Adaptive learning | High computational cost | | Blockchain [7] | Tamper-proof logs | Network overhead | Recent works focus on: - Context-aware authentication (CAA) for ECUs [8] - Hardware security modules (HSMs) in gateways [9] - Federated learning for collaborative detection [10] ## III. System Under Analysis *Three-layer architecture showing attack surfaces in telematics, CAN bus, and ECU communication* ## IV. Existing Solutions ### A. Cryptographic Protections - **AES-128** for CAN payload encryption [1][2] - **HMAC-SHA256** for message authentication [5] - **ECU-specific keys** using HSMs [9] ### B. Anomaly Detection - LSTM networks for temporal pattern analysis [6] - Random forest classifiers for attack identification [10] ## V. Discussion **Tradeoffs Identified:** - 12ms latency increase with HMAC authentication - 18% false positives in LSTM detection - 45% ECU memory overhead for key storage **Recommendations:** 1. Hybrid approach combining ML and lightweight crypto 2. Hardware-assisted key management 3. Standardized security testing framework ## VI. Conclusion Our analysis reveals critical gaps in current automotive security architectures. The proposed framework reduces attack surface by 62% in simulated environments while maintaining real-time performance requirements. Future work will focus on post-quantum algorithms for V2X communications. ## References [1] Author et al., "CAN Security Survey," IEEE Trans. Veh. Tech., 2023 [2] Researcher et al., "ECU Authentication," AutoSec Conf., 2024 ... (10+ references following IEEE format)