Privacy-Balanced Protection Methodology for Vehicle Health Monitoring Under the Federated Learning Framework

In the past few years, with the increasing popularity of new energy vehicles, the social concern for vehicle health monitoring has also been growing. Traditionally, vehicle health monitoring mainly relies on manual inspections and regular maintenance, which are inefficient and costly. With the rapid development of the Internet of Vehicles (IoV), vehicle health monitoring has become more common and feasible in internet-connected vehicles. In this trend, this paper proposes a Federated Learning-based vehicle health monitoring scheme, which is applied to connected vehicles in the IoV. To protect the privacy rights of vehicle owners, we propose the Federated Vehicle Model Training algorithm (FedVmt) in this scheme, which trains and aggregates models locally rather than on a central server. During the training process, the central server is responsible for distributing encrypted public keys and decrypting encrypted data, ensuring the security of the data. At the same time, we also use homomorphic encryption technology to further ensure the security of the data. Homomorphic encryption technology allows for computational operations to be performed in an encrypted state without the need to decrypt the data, effectively preventing potential risks during data transmission and processing.