Vehicular Blockchain-Based Collective Learning for Connected and Autonomous Vehicles
【Author】 Fu, Yuchuan; Yu, Fei Richard; Li, Changle; Luan, Tom H.; Zhang, Yao
【Source】IEEE WIRELESS COMMUNICATIONS
【影响因子】12.777
【Abstract】The accuracy of the ML model is essential for the further development of AI-enabled CAVs. With the increasing complexity of on-board sensor systems, the large amount of raw data available for learning can however cause big communication burdens and data security issues. To alleviate the communication cost yet improve the accuracy of machine learning with preserved data privacy is an important issue to address in CAVs. In this article, we survey the existing literature toward efficient and secured learning in a dynamic wireless environment. In particular, a BCL framework for AI-enabled CAVs is presented. The framework enables distributed CAVs to train ML models locally and upload to blockchain network to overall utilize the "collective intelligence" of CAVs while avoiding large amounts of data transmission. Blockchain is then applied to protect the distributed learned models. We evaluate the performance of the presented framework by simulations and discuss a range of open research issues that need to be addressed in the future.
【Keywords】Data models; Computational modeling; Servers; Data privacy; Machine learning
【发表时间】2020 APR
【收录时间】2022-01-02
【文献类型】
【主题类别】
--
【DOI】 10.1109/MNET.001.1900310
评论