Application Research on Blockchain-based Asynchronous Federated Learning in Data Security Sharing in Vehicle Network

In the vehicle-connected environment of intelligent transportation systems, although data sharing brings new development opportunities, it is also accompanied by challenges such as privacy and security (such as data leaks) and upload efficiency. These issues reduce the enthusiasm of data providers to participate. In order to solve the above problems, this research aims to propose a new architecture based on asynchronous federated learning, focusing on solving bandwidth, security and privacy issues encountered in IoT vehicle data sharing, and is committed to improving the reliability of data sharing. and efficiency. By adopting blockchain technology to enhance the security and reliability of model parameters, reducing transmission loads and alleviating privacy concerns. At the same time, the research also uses a distributed, efficient and secure data sharing architecture based on blockchain to select nodes and further optimizes the asynchronous federated learning scheme to improve overall efficiency. Experimental results show that the proposed scheme not only improves the learning accuracy, but also significantly accelerates the convergence speed, and at the same time enhances privacy protection and service quality, fully verifying the effectiveness and practicality of this research method.