Blockchain-based federated learning for internet of Vehicles

The purpose of this study is to construct a block chain encryption federation learning model for vehicle networking scenario, in order to solve the problems of data security, privacy protection and model training efficiency in the pre-test scenario of vehicle networking traffic. The model integrates federation learning, block chain and homomorphic encryption technology, and designs a roadside unit federation learning framework based on block chain, which ensures the integrity and nonrepudiation of data transmission and solves the problem of data tampering in the process of communication. At the same time, homomorphic encryption algorithm is introduced to ensure the safe transmission of model parameters and avoid gradient leakage. Finally, the use of small batch gradient descent algorithm reduces the communication overhead and improves the training efficiency. The experimental results show that the model improves the security and efficiency of the data sharing scheme, and successfully achieves the purpose of traffic prediction.