DoS攻击下多智能体系统无模型自适应迭代学习跟踪控制

Model free adaptive iterative learning tracking control for multi-agent systems under DoS attacks

中文关键词  无模型自适应控制  迭代学习控制  数据量化  周期性DoS攻击  多智能体系统

英文关键词  model free adaptive control  iterative learning control  data quantization  periodic DoS attacks  multi-agent systems

中文摘要

      针对周期性拒绝服务(DoS)攻击下多智能体系统有限时间趋同跟踪控制问题, 本文提出了一种无模型自适 应迭代学习控制(MFAILC)算法. 假设多智能体系统具有固定拓扑结构, 并且仅有部分智能体可获取到期望轨迹信 息. 在多智能体系统数据传输过程中, 需要经由对数量化器进行量化处理. 首先, 使用伪偏导数将智能体系统动态 线性化, 处理过程中考虑符合伯努利分布的周期性DoS攻击现象, 在此基础上设计了MFAILC控制算法, 其次, 采用 压缩映射方法给出了一个在期望意义下保证跟踪误差收敛的充分条件, 并在理论上证明了所提算法的收敛性. 所 提算法只需利用系统的输入输出数据就可完成趋同跟踪任务. 最后, 仿真结果验证了所提算法的有效性.

英文摘要

      A model free adaptive iterative learning control (MFAILC) algorithm is proposed for the finite-time convergence-tracking control problem of multi-agent systems under periodic Denial-of-Service (DoS) attacks. It is assumed that the multi-agent system has a fixed topology, and only some agents can obtain the desired trajectory information. In the process of multi-agent system data transmission, it needs to be quantized by logarithmic quantizer. Firstly, the pseudo partial derivative is used to linearize the agent system dynamically, and the periodic DoS attack in accordance with Bernoulli distribution is considered in the process of processing. On this basis, the MFAILC scheme is designed. Then, a sufficient condition to ensure the convergence of tracking error in the expectation sense is given by using the approach of contraction mapping, and the convergence of the proposed algorithm is proved theoretically. The proposed algorithm can complete the convergence tracking task only by using the input and output data of the system. Finally, the simulation results verify the effectiveness of the proposed algorithm.