A secure and efficient blockchain enabled federated Q-learning model for vehicular Ad-hoc networks

Area of Science:

• Computer Science
• Network Security
• Artificial Intelligence

Background:

• Vehicular Ad-hoc Networks (VANETs) face increasing security threats due to the rise of automated vehicles.
• Existing Federated Learning (FL) methods offer limited privacy protection for VANET data exchange.
• Secure data transfer is critical for maintaining VANET integrity.

Purpose of the Study:

• To propose a novel security framework for VANETs by merging Blockchain technology with Federated Q-learning.
• To enhance data security and privacy in VANETs through advanced cryptographic and storage techniques.
• To improve the overall safety and performance of VANETs.

Main Methods:

• Implemented Extended Elliptic Curve Cryptography (EX-ECC) for data encryption.
• Utilized Federated Q-learning for privacy-preserving data training.
• Integrated Interplanetary File System (IPFS) for secure Blockchain storage.
• Employed Delegated Practical Byzantine Fault Tolerance (DPBFT) for network validation.

Main Results:

• The proposed framework significantly enhances VANET security and privacy.
• Achieved high throughput (102465.8 KB/s) and low communication overhead (360.57 Mb).
• Demonstrated superior performance compared to existing VANET security approaches.

Conclusions:

• The integration of Blockchain with Federated Q-learning offers a robust solution for VANET security.
• The proposed EX-ECC and DPBFT-based framework effectively addresses privacy and security concerns.
• This approach holds significant potential for advancing the safety and efficiency of intelligent transportation systems.