A Data Grading Privacy Preservation Method for Intelligent Connected Vehicle

During the design, development, and operational phases of intelligent connected vehicles (ICVs), an increased amount of data is required to be published. This raises privacy concerns for drivers, as ICV data contain rich sensitive information, such as mileage, battery energy information, and fuel level. To address these concerns, local differential privacy techniques have been developed. These techniques aim to balance data utility and privacy preservation and have shown positive results in the big data domain. Building on these techniques, this paper introduces a data grading privacy preservation method to enhance the privacy preservation of vehicle data published. This method employs the $\mathbf{k}$-anonymity mathematical framework and information entropy to grade data attributes into different levels. It then integrates the sensitivity into the differential privacy budget. It achieves adaptive privacy preservation of data attributes in multiple aspects. Differential privacy based on the Laplace noise mechanism is employed to safeguard data security. Experiments show that the method proposed in this paper excels in reducing error and re-identification risk.