Research On Data Security Sharing Solutions For Smart Transportatio

Intelligent transportation has entered a new phase of "data-driven" with the introduction of new digital technology and applications.Unfortunately,the existing transportation sector lacks system synergy and efficient and timely information interchange,and it also lacks effective data integration and cooperation mechanisms,making meeting the demands of modern transportation development challenging.It is an inevitable trend in the development of intelligent transportation to overcome the old data silo conundrum through data sharing and achieve information integration.Data sharing across systems has serious trust and security difficulties due to the increasing volume of data and more complicated network topology in the intelligent transportation system.The primary challenges are:(1)It is challenging to strike a balance between security,integrity,and availability.(2)The balance between flexible data access and sensitive data confidentiality can be challenging.As a result,based on federal learning and Ciphertext Policy Attribute-based Encryption(CP-ABE),a data security sharing scheme for intelligent transportation is investigated from two perspectives: on the one hand,using federated learning technology to achieve secure sharing of traffic image information;on the other hand,CP-ABE is used to achieve secure sharing of traffic text data containing personal privacy information.The specific work is as follows:(1)A secure traffic image data sharing scheme based on federated learning is designed.A trustworthy multi-source data-sharing mechanism is first built.Via federated learning,data owners acquire trained machine learning models,and the blockchain facilitates federated learning by allowing model parameter exchange,minimizing reliance on individual nodes.The model lessens communication overhead while protecting the participants’ privacy to some extent.Moreover,an adaptive differential privacy technique and a model quality-based incentive mechanism for verification are designed.A solid balance between upholding a privacy budget and model accuracy,results in an increased initiative of users to participate in the system.Experimental results show that,in addition to obtaining high model accuracy,the approach also performs well in reducing privacy budget and withstanding harmful attacks.(2)A secure traffic text data sharing scheme based on CP-ABE is designed.This study begins by constructing a searchable data-safe sharing model.Sensitive data sets are pre-processed with the(p,ε,k)-anonymity data desensitization technique before encryption to minimize their sensitivity.Based on this,a searchable attribute encryption approach is employed to accomplish multi-keyword search and increase search efficiency by utilizing blockchain benefits such as traceability.This program provides the data owner the power to handle data access while also allowing data requesters to construct trapdoors depending on their interests and submit query requests to the data owner,ensuring flexible data access control.The theoretical and experimental analysis demonstrates that the scheme can not only realize fine-grained access and many-to-many matching of data,thus achieving the purpose of cross-domain data sharing among multiple systems,but also effectively guarantee the privacy and information security of users.