A Framework for Privacy-Preserving in IoV Using Federated Learning With Differential Privacy
Vehicles become more advanced and smarter due to advancements in technology in the modern world. Every person now a days, demand a smart vehicle due to their automobility and smart controls. This is all possible through advancements in VANET (Vehicular Adhoc Network) and the Internet of Vehicles (Io...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10834602/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Vehicles become more advanced and smarter due to advancements in technology in the modern world. Every person now a days, demand a smart vehicle due to their automobility and smart controls. This is all possible through advancements in VANET (Vehicular Adhoc Network) and the Internet of Vehicles (IoV). Vehicles in the VANET are highly connected to each other and this thing can cause security, safety, and privacy risks for the asset itself and driver also. It can become a reason of major threat. And these threats can occur due to tracing the location of the vehicle. Existing techniques like group-based shadowing schemes, obfuscation, silent periods, and mix-zone have preserved privacy of location somehow, but don’t have a good QoS and optimized efficient security. To overcome these issues, we introduced a new privacy framework, which is an improvement of the existing shadowing scheme. We proposed a computationally efficient group leader selection process based on centeredness, rule obeyed, and OBU resources, reducing overhead by 20%, integrating FL with DP to preserve data privacy without sacrificing utility, and achieving a 15% improvement in location accuracy under privacy constraints, validating the scalability and robustness of the framework through extensive simulations involving up to 300 vehicles. Group Leader is used as an optimization of the overall framework including efficiency and implementation of the scheme. This scheme increases privacy if the number of vehicles also increases, and this thing makes our scheme more scalable. This scheme overcomes the many drawbacks of existing techniques like a higher tracing ratio in shadowing schemes, totally depending on the group leader, and reduced utility of all schemes based on distances. The most important thing, the single point of failure in the group leader base shadowing scheme is overcome by using local federated learning with differential privacy. Validation results of our proposed scheme showed that it outperformed the current schemes mainly based on group leader. |
|---|---|
| ISSN: | 2169-3536 |