基于联邦学习和区块链的车联网隐私保护数据共享方案  

Privacy-preserving data sharing scheme for Internet of Vehicles based on federated learning and blockchain

作　　者：王杨鹏 牛宪华 熊玲 陈鹏 徐雷[2] WANG Yangpeng;NIU Xianhua;XIONG Ling;CHEN Peng;XU Lei(School of Computer and Software Engineering,Xihua University,Chengdu Sichuan 610039,China;School of Emergency Management,Xihua University,Chengdu Sichuan 610039,China)

机构地区：[1]西华大学计算机与软件工程学院,成都610039 [2]西华大学应急管理学院,成都610039

出　　处：《计算机应用》2024年第S01期107-111,共5页journal of Computer Applications

基　　金：国家自然科学基金资助项目(62171387,62202390);四川省自然科学基金资助项目(2022NSFSC0556)。

摘　　要：车联网(IoV)路况监测需对用户隐私数据进行传输、存储与分析等处理,因此保障用户隐私尤为重要。基于联邦学习与区块链技术,提出一种面向IoV的分层联邦学习框架,并基于所提框架设计一种轻量级的隐私保护数据共享方案。所提方案采用掩码技术加密车辆共享的模型参数,保障参数在共享过程中的机密性;同时,引入基于模型参数准确性的轻量级共识机制和审计机制,提高数据共享的安全性和可靠性。仿真实验结果显示,所提方案模型精度相较于其他方案提高了5%,同时在恶意环境下模型精度相较于未加入审计机制的方案提高了30%。Monitoring road conditions in Internet of Vehicles(IoV)involves processing of user privacy data,including transmission,storage,and analysis.Therefore,protecting user privacy is critical.Based on federated learning and blockchain technology,a hierarchical federated learning framework for the IoV was proposed,and a lightweight privacy-preserving data sharing scheme was designed based on the proposed framework.Mask technology was used in the proposed scheme to encrypt the model parameters of vehicle sharing to guarantee the confidentiality of the parameters in the sharing process;at the same time,a lightweight consensus mechanism and an auditing mechanism based on the accuracy of the model parameters were introduced to improve the security and reliability of data sharing.The simulation results demonstrate that,the proposed scheme increases the model accuracy without attackers by 5%compared to other schemes,and improves the model accuracy by 30%compared to the schemes without auditing mechanism in the presence of attackers.

关 键 词：车联网 数据共享 隐私保护 联邦学习 区块链 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]