A blockchain-based decentralized machine learning framework for collaborative intrusion detection within UAVs

作     者：Khan, Ammar Ahmed Khan, Muhammad Mubashir Khan, Kashif Mehboob Arshad, Junaid Ahmad, Farhan 

作者机构：NED Univ Engn & Technol Karachi Dept Comp Sci & IT Karachi Pakistan NED Univ Engn & Technol Karachi Dept Software Engn Karachi Pakistan Birmingham City Univ Sch Comp & Digital Technol Birmingham W Midlands England Coventry Univ Inst Future Transport & Cities Syst Secur Grp Coventry W Midlands England 

出 版 物：《COMPUTER NETWORKS》 (Comput. Networks)

年 卷 期：2021年第196卷

核心收录：

学科分类：0810[工学-信息与通信工程] 0808[工学-电气工程] 08[工学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

主　　题：Unmanned aerial vehicles UAV Blockchain Machine learning Decentralized machine learning Collaborative intrusion detection 

摘      要：UAVs have numerous emerging applications in various domains of life. However, it is extremely challenging to gain the required level of public acceptance of UAVs without proving safety and security for human life. Conventional UAVs mostly depend upon the centralized server to perform data processing with complex machine learning algorithms. In fact, all the conventional cyber attacks are applicable on the transmission and storage of data in UAVs. While their impact is extremely serious because UAVs are highly dependent on smart systems that extensively utilize machine learning techniques in order to take decisions in human absence. In this regard, we propose to enhance the performance of UAVs with a decentralized machine learning framework based on blockchain. The proposed framework has the potential to significantly enhance the integrity and storage of data for intelligent decision making among multiple UAVs. We present the use of blockchain to achieve decentralized predictive analytics and present a framework that can successfully apply and share machine learning models in a decentralized manner. We evaluate our system using collaborative intrusion detection as a case-study in order to highlight the feasibility and effectiveness of using blockchain based decentralized machine learning approach in UAVs and other similar applications.