[18]刘小垒.基于群体智能进化算法的对抗样本生成研究 [D].成都：电子科技大学 (博士学位论文), 2019.

Liu X L.Research on generation of adversarial examples based on swarm evolutionary algorithm [D].Chengdu: School of Information and Software Engineering (Doctoral dissertation), 2019.

[19]伍益明.恶意攻击下的多智能体系统安全一致性问题研究 [D].杭州: 浙江工业大学 (博士学位论文), 2016.

Wu Y M.Research on secure consensus for multi-agent systems under malicious attacks [D].Hangzhou: Zhejiang University of Technology (Doctoral dissertation), 2016.

[20]崔朝阳, 孙甲琦, 徐松艳, 等.适用于集群无人机的自组网安全分簇算法 [J].山东大学学报 (理学版), 2018, 53(7): 51-59.

Cui Z Y, Sun J Q, Xu S Y, et al.A secure clustering algorithm of Ad Hoc network for colony UAVs [J].Journal of Shandong University (Natural Science), 2018, 53(7): 51-59.

[21]臧义华, 李小娟.基于区块链的无人集群作战信息共享架构 [J].指挥控制与仿真, 2020, 42(4): 19-22.

Zang Y H, Li X J.Unmanned cluster operation information sharing [J].Command Control & Simulation, 2020, 42(4): 19-22.

[22]范泉涌, 金茂菁, 黄玲, 等.无人驾驶汽车的安全问题及对策 [J].科技中国, 2019 (6): 13-15.

Fan Q Y, Jin M J, Huang L, et al.The safety problems and countermeasures of unmanned vehicles [J].Scitech in China, 2019 (6): 13-15.

[23]李修全.人工智能应用中的安全、隐私和伦理挑战及应对思考 [J].科技导报, 2017, 35(15): 11-12.

Li X Q.Security, privacy and ethical challenges in artificial intelligence applications and their countermeasures [J].Science & Technology Review, 2017, 35(15): 11-12.

[24]Johnson J.Artificial intelligence, drone swarming and escalation risks in future warfare [J].The RUSI Journal, 2020, 165(2): 26-36.

Artificial Intelligence Security in Multiple-Unmanned System Cooperation

Shi Wen 1, Wang Kaiwen1, Yu Chengpu 1, Sun Jian 1, Chen Jie 1, 2

(1.School of Automation, Beijing Institute of Technology, Beijing 100081, China; 2.Tongji University, Shanghai 200092, China)

Abstract: Multiple unmanned system cooperation is an important component of China’s new-generation artificial intelligence (AI) planning and it is a transformative technology for future national defense construction and social development of China.Although significant achievements have been made regarding the technological research and system integration of multiple unmanned system cooperation, research on AI security is still in its infancy.Promoting multiple unmanned system cooperation enables AI application and risk control.In this article, we propose a four-in-one strategy to promote the collaborative and secure development of multiple unmanned systems, investigate potential challenges and countermeasures for multiple unmanned system cooperation at the endogenous and derivative security levels, and propose several suggestions regarding the security of intelligent unmanned systems.Specifically, a national unmanned system verification platform should be established to promote the construction of talent teams; services need to be upgraded for the unmanned system industry to develop a new-generation AI security ecology; and the advantages of multiple unmanned system cooperation should be maximized to improve people’s livelihood.

Keywords: multiple unmanned system cooperation; artificial intelligence (AI) security; security risk prevention and control

中图分类号：TP18

文献标识码：A

DOI 10.15302/J-SSCAE-2021.03.007

收稿日期：2021-02-15；

修回日期：2021-04-29

通讯作者：俞成浦，北京理工大学教授，主要研究方向为系统辨识与机器学习、分布式优化与控制、传感器网络与室内定位；E-mail: yuchengpu@bit.edu.cn

资助项目：中国工程院咨询项目“新一代人工智能安全与自主可控发展战略研究”(2019-ZD-01)

本刊网址：www.engineering.org.cn/ch/journal/sscae