Federated Learning on Wheels: A Decentralized Approach to Privacy-Enhanced Data Collection in Internet of Vehicles

Due to privacy issues and the scattered nature of data produced by vehicles, the Internet of Vehicles (IOV) poses considerable hurdles for data collecting. In this chapter, we examine the idea of “Federated Learning on Wheels” (FLoW), which provides a decentralised method for IOV data collection with a focus on privacy. FLOW makes use of the onboard computer resources of cars to carry out model training locally, making sure that private information stays on the cars and is not shared with a centralised server. This strategy overcomes the shortcomings of conventional centralised data collecting approaches while simultaneously protecting user privacy. We examine the fundamentals of federated learning and how they relate to IOV, highlighting the advantages of maintaining privacy. We also look at secure aggregation procedures and confidentiality safeguards as additional methods for privacy-enhanced data acquisition in FLOW. Additionally, we emphasise the significance of accuracy and performance issues in decentralised contexts and use examples that illustrate FLOW's usefulness. We also explore security and trust issues, talking about possible weaknesses and methods to secure the reliability of participants and model updates. We also consider how blockchain technology may be incorporated for improved security and openness. We conclude by discussing FLOW future directions, difficulties, and ethical issues in order to shed light on its possible significance and legal ramifications. Overall, this chapter clarifies the relevance of Federated Learning to Wheels as a ground-breaking approach to data collecting with increased privacy in the Internet of Vehicles.

Keywords: Blockchain, Federated learning, Internet of vehicles (IOV), Privacy preservation.