基于自适应动态规划的时滞多智能体系统协同输出调节  

Cooperative Output Regulation of Multi-agent Systems with Time Delay Based on Adaptive Dynamic Programming

作　　者：赵纯彦 徐磊 高伟男 杨涛[1] ZHAO Chunyan;XU Lei;GAO Weinan;YANG Tao(College of Information Science and Engineering,Northeastern University,Shenyang 110819,China)

机构地区：[1]东北大学信息科学与工程学院,辽宁沈阳110819

出　　处：《控制工程》2023年第8期1477-1487,共11页Control Engineering of China

基　　金：国家自然科学基金重大项目(61991403)。

摘　　要：研究一类在状态、控制输入上具有多个时滞的离散时间多智能体系统的协同输出调节问题。首先,引入了一种扩展状态的方法,在不需要知道时滞的情况下,将有时滞的系统转化为扩展的无时滞系统;然后,利用内模原理和值迭代相结合的分布式控制方法来实现该类问题的最优控制;其次,在不依赖系统动力学知识和初始稳定控制策略的情况下,仅使用系统的控制输入和状态数据,在线学习出一种基于值迭代的数据驱动的自适应控制策略,能够保证闭环多智能体系统既能抑制干扰又能实现渐近跟踪;最后,通过一个仿真案例验证了所提方法的有效性。The problem of cooperative output regulation for discrete-time multi-agent systems with multiple delays in the state and control input is studied.Firstly,an extended state-space augmentation approach is used,which brings the system into a delay-free form without requiring knowledge of the delays.Then,the distributed control method that combines the internal model principle and the value iteration method is proposed to achieve cooperative output regulation in an optimal sense.Secondly,a data-driven adaptive control strategy based on value iteration is learned online using only the control input and state data of the system without relying on the knowledge of the system dynamics and the initial stabilizing control strategy.This ensures that closed-loop multi-agent systems not only suppress interference but also realize asymptotic tracking.Finally,an example is used to demonstrate the effectiveness of the methods proposed in our study.

关 键 词：多智能体系统 输入延时 状态延时 协同输出调节 值迭代 自适应动态规划 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]