具有执行器故障的不确定多智能体系统自适应动态面控制

本文研究了一类具有输入量化、未建模动态和执行器故障的非线性多智能体系统的一致跟踪问题.引入一个可量测的动态信号消除未建模动态对系统的影响.利用Young’s不等式和高斯函数的性质,有效地处理了多智能体邻居节点在设计的第一步中对子系统的耦合作用.通过将滞回量化器表示为具有有界系数和有界扰动的输入线性函数,并利用动态面控制方法,提出一种自适应神经网络动态面控制方案,简化了控制器的设计,保证了闭环系统的所有信号都是半全局一致终结有界的,所有跟随者都能实现期望的一致性.最后,仿真结果验证了所提出的自适应控制策略的有效性.     

重载列车多智能体模型的鲁棒一致性控制方法

重载列车全长数公里,其运行过程是复杂的动力学系统.重载列车自动驾驶的关键核心技术是跟踪给定的速度曲线.以重载列车智能货车方案为基础,通过分析列车运行动力学过程,建立重载列车多智能体模型;考虑列车运行时外界的未知干扰,同时保证车厢间处于安全距离,提出一种重载列车复合一致性控制器:用相邻车厢单元的速度等信息构建一致性算法并引入滑模控制加快系统速度一致性收敛;列车不同车厢受到干扰视为未知扰动,且随着滑模增益增加会使系统抖动较大、鲁棒性削弱,所以设计观测器估计扰动并补偿至控制器保证系统收敛并提高抗干扰性;引入人工势函数确保相邻车厢单元间距处于安全范围内,减小纵向冲动.采用Matlab软件进行仿真,跟踪给定速度曲线,并用多种干扰来模拟未知因素对列车的影响,与不加观测器的控制器效果进行对比.仿真结果表明:该复合一致性控制器能够较好跟踪设定速度曲线,速度误差保持在±0.4 km/h以内,且车厢间距处于设定的安全范围内;与不加观测器的控制器在同样干扰下作对比,所提出的控制器的速度跟踪误差仍然在±4 km/h以内,且对于未知干扰有较好的鲁棒性.     

非完整约束多智能体系统基于屏障控制函数的分布式协同控制

本文设计了一种基于屏障控制函数(CBF)的分布式协同控制算法,实现了领航–跟随者框架下非完整约束多智能体系统的连通性与编队控制.首先,通过将连通性保持问题建模为系统约束,定义了该约束的调零屏障函数(ZBF).其次,在此基础上,构建李雅普诺夫函数与角速度输入之间的关系,对跟随者智能体设计了基于调零屏障函数的协同控制算法,其中线速度控制器保证跟随者的速度的跟踪与队形的跟踪,而梯度型角速度控制器实现跟随者角度的矫正.然后,利用调零屏障函数不变集相关引理证明了连通性约束集为正不变集,若初始时刻连通,则跟随者智能体始终与领航者保持连通性.同时,本文提出的算法实现编队误差的渐近收敛.本文中的队形适用常见的固定队形编队需求,也适用于领航者是动态(有线速度和角速度)的情况.最后,通过数值仿真进一步验证了该算法在不同队形需求下的有效性.     

多智能体系统的有限时间包含控制研究

在固定有向拓扑结构下,研究了具有多个静态或动态领导者的多智能体有限时间包含控制问题;假设领导者之间不存在信息的交互,提出基于快速终端滑模的控制算法,该算法驱使跟随者的运动轨迹在有限时间收敛到由领导者组成的凸包中;进一步,考虑实际应用中跟随者状态不能在线获得的情况,提出基于有限时间观测器的包含控制协议;利用图论和Lyapunov有限时间稳定性理论,给出了有限时间包含控制的充分条件;最后通过仿真示例,验证了算法的有效性.     

基于多智能体强化学习的无人艇协同围捕方法

针对多无人艇对海上逃逸目标的围捕问题,提出一种基于多智能体强化学习的围捕算法.首先,以无人艇协同进攻为背景建立无边界围捕问题的环境和运动学模型,并针对快速性和合围性的需求给出围捕成功的判定条件;然后,基于多智能体近端策略优化(MAPPO)算法建立马尔可夫决策过程框架,结合围捕任务需求分别设计兼具伸缩性和排列不变性的状态空间,围捕距离、方位解耦的动作空间,捕获奖励与步长奖励相结合的奖励函数;最后,采用集中式训练、分布式执行的架构完成对围捕策略的训练,训练时采用课程式学习训练技巧,无人艇群共享相同的策略并独立执行动作.仿真实验表明,在无人艇起始数量不同的测试条件下,所提出方法在围捕成功率和时效性上相较于其他算法更具优势.此外,当无人艇节点损毁时,剩余无人艇仍然具备继续执行围捕任务的能力,所提出方法鲁棒性强,具有在真实环境中部署应用的潜力.     

基于观测器的多智能体系统的有限时间包含控制

在固定有向拓扑结构下,研究了具有多个静态或动态领导者的多智能体有限时间包含控制问题;假设领导者之间不存在信息的交互,提出基于快速终端滑模的控制算法,该算法驱使跟随者的运动轨迹在有限时间收敛到由领导者组成的凸包中;进一步,考虑实际应用中跟随者状态不能在线获得的情况,提出基于有限时间观测器的包含控制协议;利用图论和Lyapunov有限时间稳定性理论,给出了有限时间包含控制的充分条件;最后通过仿真示例,验证了算法的有效性。     

随机多智能体系统的有限时间多目标环绕控制

针对随机多智能体系统,研究了其多目标有限时间环绕控制问题,其中每个智能体可以跟踪多个目标,但是每个目标只能被一个智能体跟踪.首先,为每个智能体设计了目标状态的估计器,证明了在白噪声干扰之下每个智能体都可以在有限时间内对目标的几何中心位置作出估计.其次,为每个智能体设计了环绕半径的估计器,以及分布式环绕控制器.然后,利用李雅普诺夫有限时间随机稳定性理论,证明了所有智能体在有限时间内能够对环绕半径作出估计,同时可以实现有限时间环绕控制.最后,通过仿真验证了结果的有效性和正确性.     

基于一类连续非线性函数的多智能体系统有限时间一致性

针对存在时滞的多智能体系统,提出了基于一类连续非线性函数的有限时间一致性算法.利用Lyapunov有限时间稳定性理论和矩阵理论,给出了这类算法使得系统能够在有限时间内达到一致的充分条件,进而给出了一个满足条件的有限时间一致性算法,并对该算法的收敛性进行分析,得到了系统的收敛时间.数值仿真验证了所提出算法的有效性.     

带输出死区的多智能体系统预设时间事件触发式协同控制

针对带有输出死区的多智能体系统的一致性控制问题,提出一种可预设时间的事件触发式协同控制方法.输出死区现象普遍存在于实际系统,对控制回路中负反馈调节影响较大,会导致系统控制性能的下降.为此,通过结合Nussbaum函数对输出死区特性进行补偿,以削弱其对系统性能的影响,但上述问题解决的同时也会加剧对系统控制传输资源的占用.由于多智能体系统依靠智能体之间频繁的信息交互实现控制目标,自身传输资源有限.考虑到实际系统的限制,利用事件触发控制策略节省系统的控制传输资源.更进一步,为有效提高系统性能,使系统能够快速达到稳定,引入一类转换函数进行系统变换,实现系统同步误差在预设时间内收敛于紧集的目标.理论分析和仿真结果验证了所提出方法的有效性.     

具有随机时延的的多智能体系统的一致性

本文研究了具有随机时不变通信时延的多智能体系统的一致性问题。假设系统拓扑是固定有向的强连通图,考虑多智能体系统中存在随机通信时延的情况,即各智能体之间的通信时延是以一定概率存在的,采用具有动态参考状态的一致性控制算法,通过变量代换,将多智能体的一致性问题转化为误差系统的渐进稳定性问题,然后再应用Lyapunov稳定性理论与随机分析方法推导出误差系统在该算法作用下趋于稳定的充分条件,并以线性矩阵不等式（LMIs）表示,即原多智能体系统在此充分条件下各智能体的状态能够达到一致。最后应用一个实例验证了所提出结论的有效性。     

随机激励下板球系统建模与有限时间全状态预设性能跟踪控制

研究板球系统受到随机激励时的数学建模与轨迹跟踪控制问题.首次建立了板球系统的随机数学模型,并结合backstepping方法、有限时间预设性能函数、全状态约束及新的预设性能推导方法设计了具有未知输入饱和的随机板球系统实际有限时间全状态预设性能跟踪控制器,实现了随机激励下板球系统的有限时间预设性能轨迹跟踪控制.所设计的控制器保证了系统跟踪误差能够被预先给定的有限时间性能函数约束,并且能在任意给定的停息时间内收敛到预先给定的邻域内.最后通过仿真实验验证了所设计控制器具有更好的控制效果.     

Leader-follower finite-time formation control of multiple quadrotors with prescribed performance

This paper investigates the leader-follower formation problem for a group of quadrotors. Finite-time control scheme and prescribed performance control method, which are regarded as two highlights, are introduced in this paper. First, a control scheme with prescribed performance is used to control the translational movements to ensure the quadrotors obtain a relative gentle transient process and an adjustable steady-state error bound. Then, the desired orientations for the rotation subsystem provided by translational movements part are stabilised by a fixed-time control law. Finally, by designing a finite-time formation controller, followers can track the desired position and heading angle in finite time, which is important for the practical application. Several simulation results are given to show the effectiveness of the designed control strategy.     

Transverse function control with prescribed performance guarantees for underactuated marine surface vehicles

This paper studies the problem of stabilizing reference trajectories (also called as the trajectory tracking problem) for underactuated marine vehicles under predefined tracking error constraints. The boundary functions of the predefined constraints are asymmetric and time‐varying. The time‐varying boundary functions allow us to quantify prescribed performance of tracking errors on both transient and steady‐state stages. To overcome difficulties raised by underactuation and nonzero off‐diagonal terms in the system matrices, we develop a novel transverse function control approach to introduce an additional control input in backstepping procedure. This approach provides practical stabilization of any smooth reference trajectory, whether this trajectory is feasible or not. By practical stabilization, we mean that the tracking errors of vehicle position and orientation converge to a small neighborhood of zero. With the introduction of an error transformation function, we construct an inverse‐hyperbolic‐tangent‐like barrier Lyapunov function to show practical stability of the closed‐loop systems with prescribed transient and steady‐state performances. To deal with unmodeled dynamic uncertainties and external disturbances, we employ neural network (NN) approximators to estimate uncertain dynamics and present disturbance observers to estimate unknown disturbances. Subsequently, we develop adaptive control, based on NN approximators and disturbance estimates, that guarantees the prescribed performance of tracking errors during the transient stage of on‐line NN weight adaptations and disturbance estimates. Simulation results show the performance of the proposed tracking control.     

基于预设性能的非仿射非线性系统自适应有限时间控制

针对一类具有死区的非仿射非线性系统,将预设性能控制与有限时间控制相结合,提出一种具有预设性能的自适应有限时间跟踪控制方法.基于Backstepping技术、模糊逻辑系统及有限时间Lyapunov稳定理论,给出使系统半全局实际有限时间稳定(semi-globally practically finite-time stable,SGPFS)的充分条件和设计步骤.该控制策略不仅使系统的输出误差在有限时间内收敛到一个预先设定区域,同时保证其收敛速度、最大超调量和稳态误差均满足预先设定的性能要求.最后通过仿真示例验证了所提出设计方法的有效性.     

伺服电机驱动的连铸结晶器振动系统受限状态下预设性能控制

以伺服电机驱动的连铸结晶器振动位移系统为研究对象,针对系统输入和状态受限问题,考虑系统存在的参数不确定性和负载转矩扰动影响,设计一种基于观测器的预设性能自适应控制器.首先,针对系统存在的参数不确定性、负载转矩扰动等问题,基于Lyapunov函数设计变增益扩张状态观测器,在保证观测精度的同时,削弱峰值现象;其次,考虑状态和输入受限的情况,将预设性能函数与Backstepping技术相结合设计控制器,构建指令滤波器解决“计算膨胀”问题,引入动态补偿量对观测器及受限状态产生的误差进行补偿,并对所设计的控制器进行稳定性分析;最后,通过仿真对比实验验证控制器的有效性.     

Dynamic workforce allocation in a constrained flow shop with multi-agent system

The present work addresses the problem of real time workforce scheduling in assembly lines where the number of operators is less to the number of workstations.The problem is faced developing a two-steps procedure made of (i) a centralized scheduling based on a constraint optimization problem (COP) for initial operator scheduling, and (ii) a decentralized algorithm performed by a multiagent system (MAS) to manage workers in case of unforeseen events.In the proposed MAS architecture, Agents represent the operators trying to find local assignments for themselves. The system is validated with a simulation model and implemented with a hardware infrastructure in a real assembly line of electromechanical components. The main original contribution of the paper consists in proving – by means of both validation through a simulation model and test in a real assembly line of electromechanical components – that (1) multi-agent systems could be successfully adopted to solve a workforce scheduling problem, and (2) a combined approach consisting of centralized + distributed approach would provide better results compared with the application of one of the two approaches alone.     

一类非线性大系统分散自适应预设性能有限时间跟踪控制

研究一类非线性互联大系统的分散自适应预设性能有限时间跟踪控制问题.结合神经网络自适应技术、实际有限时间控制理论和预设性能控制方法,提出一种新的预设性能控制设计方法,以解决传统预设性能方法难以实现分散控制的问题.所设计的控制器能够保证大系统中各个子系统的跟踪误差被有限时间性能函数约束,在任意给定的停息时间内收敛到平衡点的一个给定的邻域内,且该闭环大系统的所有信号是实际有限时间稳定的.特别地,该停息时间与系统初始状态无关.两个仿真例子验证了所提出控制方法的有效性和优越性.     

Neuro-adaptive cooperative tracking control with prescribed performance of unknown higher-order nonlinear multi-agent systems

This paper is concerned with the design of a distributed cooperative synchronisation controller for a class of higher-order nonlinear multi-agent systems. The objective is to achieve synchronisation and satisfy a predefined time-based performance. Dynamics of the agents (also called the nodes) are assumed to be unknown to the controller and are estimated using neural networks. The proposed robust neuro-adaptive controller drives different states of nodes systematically to synchronise with the state of the leader node within the constraints of the prescribed performance. The nodes are connected through a weighted directed graph with a time-invariant topology. Only few nodes have access to the leader. Lyapunov-based stability proofs demonstrate that the multi-agent system is uniformly ultimately bounded stable. Highly nonlinear heterogeneous networked systems with uncertain parameters and external disturbances were used to validate the robustness and performance of the new novel approach. Simulation results considered two different examples: single-input single-output and multi-input multi-output, which demonstrate the effectiveness of the proposed controller.     

Discrete-time extended state observer-based model-free adaptive constrained sliding mode control with modified prescribed performance via CFDL and PFDL

To tackle the trajectory tracking problem and achieve high control accuracy in many actual nonlinear systems with unknown dynamics and input saturation, a novel discrete-time extended state observer-based model-free adaptive constrained sliding mode control with modified prescribed performance is investigated via compact-form dynamic linearization (CFDL) and partial-form dynamic linearization (PFDL). Firstly, the original non-affine system is turned into an affine one comprising an unknown nonlinear term and a linearly parametric term affine to the input via both PFDL and CFDL. Then, a discrete-time extended state observer (DESO) is used to estimate the lumped disturbance containing the unknown nonlinear time-varying term and the term relevant to the estimation error of pseudo partial derivative (PPD) parameter. Furthermore, a modified prescribed performance function is introduced in the model-free adaptive sliding mode control scheme to keep the output tracking error in the prescribed bound without causing any asymmetric offset error in the steady-state. Meanwhile, to suppress the influence of input saturation on the control system, an anti-windup compensator is used. Finally, rigorous theoretical analyses show the robust convergence of the tracking error via the proposed CFDL and PFDL-based methods under external disturbances. Simulations verify the superiority of the modified prescribed performance function, DESO, and anti-windup compensator in the proposed method. Also, the effects of the PFDL-based method and the CFDL-based one are compared during the simulation.