Reinforcement Learning-Based Cooperative Target Tracking Control of Unmanned Surface Vehicles Under Data Falsification Attacks
IEEE Transactions on Intelligent Vehicles(2024)
摘要
This paper deals with the problem of cooperative target tracking control for multiple networked unmanned surface vehicles under data falsification attacks. First, a data falsification attack model is constructed to account for unknown falsified control commands. Next, a networked cooperative target tracking model is developed by integrating the concurrent impacts of data falsification attacks and network-induced delays. Then, a reinforcement learning-based compensation mechanism, consisting of an actor neural network and an adaptive critic function, is developed to offset the adverse effects caused by the unknown dynamics and external disturbances. Meanwhile, to cope with data falsification attacks and reduce the neural network approximation errors, a secure control input is subtly developed. Based on the developed compensation mechanism and secure control input, a dynamic control law is constructed to regulate the action of involved unmanned surface vehicles. Furthermore, a resilient cooperative target tracking control scheme is proposed to achieve the expected cooperative target tracking control objectives. At last, two case studies are provided to validate the efficacy of the designed resilient cooperative target tracking control scheme.
更多查看译文
关键词
Unmanned surface vehicles,cooperative target tracking,reinforcement learning,data falsification attacks
AI 理解论文
溯源树
样例
![](https://originalfileserver.aminer.cn/sys/aminer/pubs/mrt_preview.jpeg)
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要