ZTID-IoV: Zero-Trust Intrusion Detection in IoV Using Neurosymbolic AI Approach with Federated Meta-Learning

The rapid growth of the Internet of Vehicles (IoVs) and smart consumer electronics has generated cybersecurity concerns that require an intelligent, adaptable, and privacy-preserving Intrusion Detection System (IDS). This study introduces ZTID-IoV, a novel neurosymbolic AI framework that integrates federated learning, lightweight transformers, and meta-learning to improve threat detection while preserving user privacy in consumer IoVs. Our approach leverages neural components such as a transformer model for recognizing patterns in network traffic, combined with symbolic AI techniques such as self-organizing maps for interpretable client clustering and rule-guided reasoning, to achieve robust cybersecurity in distributed environments. A lightweight transformer architecture optimizes performance for resource-constrained edge devices, and SOM-based clustering enhances model aggregation by grouping devices with similar behavioral patterns. The proposed system employs Model-Agnostic Meta-Learning (MAML) to enable rapid adaptation to emerging threats across diverse consumer devices, while federated learning ensures decentralized model training without exposing sensitive user data. Experiments on real-world IoT intrusion datasets demonstrate that our framework achieves higher detection accuracy compared to centralized and pure neural approaches while maintaining low computational overhead. Additionally, the neurosymbolic design provides interpretable threat explanations, crucial for consumer applications where transparency is essential. The results highlight the potential of ZTID-IoV in enabling zero-trust security for IoV and other connected consumer electronics. This work contributes to the evolving landscape of AI-driven cybersecurity by addressing critical challenges in privacy and adaptability, making ZTID-IoV particularly suitable for next-generation IoV ecosystems.