Privacy preserving data analytics in DEI empowered IoV

Abstract

The paper presents an advanced privacy-preserving data analytics framework for the Internet of Vehicles (IoV) empowered by Distributed Edge Intelligence (DEI). It integrates Federated Learning (FL), Differential Privacy (DP), and Blockchain technologies to safeguard sensitive vehicular data while enabling efficient, real-time decision-making at the network edge. FL enables decentralized model training on local vehicle data, avoiding raw data transmission and enhancing privacy. DP introduces mathematically calibrated noise to prevent individual data exposure during model updates. Blockchain provides a tamper-proof ledger for securely logging encrypted updates, ensuring data integrity and transparency. This combination minimizes latency by processing data at edge devices rather than centralized servers, reduces communication overhead, and enhances security through cryptographic protocols and consensus mechanisms. The system supports continuous learning through secure aggregation and privacy checks, making it scalable and resilient to attacks. Validated using traffic simulation data, this approach offers a robust, low-latency solution for privacy-aware analytics critical for smart, autonomous vehicular networks. This framework has the potential to streamline and expedite traffic management, as well as provide the foundation to incorporate future autonomous vehicle technology. Its modularity provides the ability to be flexible to existing IoV architectures and emergent cybersecurity challenges.