一类线性不确定变时滞系统的滑模控制

考虑一类变时滞系统的滑模控制问题,其中被控系统状态存在多个变时滞并带有不确定性.利用Lyapunov方法和Shur补引理,得到滑动模态特定的充分条件.设计一类简单的变结构控制器,使系统在有限时间内可以到达滑模面并保持在滑模面上运动.仿真结果验证本文结论的有效性.     

不确定扰动分数阶混沌系统自适应Terminal滑模同步

针对带扰动不确定分数阶混沌系统的同步问题,基于自适应Terminal滑模控制,设计了一种分数阶非奇异Terminal滑模面,保证误差系统沿着滑模面在有限时间内稳定至平衡点,在系统外部扰动和不确定性的边界事先未知的情况,设计了自适应控制率,在线估计未知边界,使得同步误差轨迹能到达滑模面。最后,以三维分数阶Chen系统和四维分数阶Lorenz超混沌系统为例,利用所设计的自适应Terminal滑模控制器进行同步仿真,验证了所给方法是有效性和可行性。     

一类不确定T-S模糊时滞系统新的滑模控制方法

针对一类时滞不确定系统,基于模糊滑模控制方法,考虑了其镇定问题.所研究系统中的时滞项和参数不确定项分别为有界时变和非匹配的.首先设计滑模面,使得系统状态在滑模面保持指数渐近稳定,并且将系统的稳定性证明转化为线性矩阵不等式的解的存在性问题;其次,基于到达条件,构造控制器,使得系统在设计控制器的作用下在有限时间内到达切换面;最后,给出仿真算例,经验证说明所设计的控制器的有效性.     

一类复杂网络同步研究的新方法

针对一类耦合网络,基于变结构控制方法研究了其同步问题。针对耦合网络设计切换面,通过极点配置,选择系统的左特征向量函数来设计切换面,使得系统的状态轨线在期望的同步轨道保持稳定;针对系统模型,设计变结构控制器。通过选取合适的到达条件构造控制器,使得从任意初始状态出发网络状态能够在有限时间内到达切换面。仿真结果表明了所设计控制器的有效性。     

时滞离散系统的非线性准滑模鲁棒控制

针对一类变时滞离散系统,基于一种新型的非线性时变准滑模面设计方案,研究了系统的鲁棒滑模控制问题.首先,给出了该种非线性切换函数的一般形式,该类准滑模面具有时变的特征,且能够动态地改善系统的运动品质.利用自由权矩阵与线性矩阵不等式技术,给出了该类时滞离散系统非线性准滑模面的设计方法,并得到了稳定的非线性准滑模面存在的充分条件;其次,基于一种改进的离散趋近律方法,设计了相应的准滑模控制器,以保证系统的状态在有限时间内到达准滑模,从而将此类非线性系统的滑模变结构控制的分析与综合问题推广到了时滞非线性系统.最后,仿真结果表明,在本文所设计的准滑模面与准滑模变结构控制器的作用下,系统的状态是稳定的,且具有响应速度快、超调量小、调节时间较小等优点,从而说明了本文所设计方法的有效性.     

非匹配的不确定时滞离散系统的滑模控制

针对一类不确定离散时滞系统,研究了其滑模控制问题。所考虑的系统是变时滞的,并且含有非匹配的时变参数不确定项。通过非奇异矩阵变换将其转化为两部分,一部分满足传统的匹配争件;另一部分是非匹配,但是满足范数有界的条件。利用变结构控制方法,设计了满足到达条件的滑模控制器,使系统状态在有限时间内到达滑模面。并且在控制器的作用下保持在滑模面上,从而使系统具有稳定的滑动模态,克服了传统滑模控制方法中不确定项需满足匹配条件的保守性。     

多关节机器人的神经滑模控制

针对含有建模误差和不确定干扰的多关节机械臂轨迹跟踪控制,提出了一种神经滑模控制方法。该方法采用全局快速终端滑模面保证了系统状态能够在有限时间内到达滑模面和平衡点。采用径向基函数神经网络自适应地补偿系统的建模误差和外界干扰,保证了滑模控制在滑模面的运动。利用李亚普诺夫稳定性判据推导出了控制器的控制律和神经网络的目标函数,通过神经网络的在线学习,削弱了滑模控制的抖振。仿真结果表明了其有效性。     

多关节机器人的模糊神经滑模控制

针对含有建模误差和不确定干扰的多关节机器人轨迹跟踪控制,提出了一种模糊神经滑模控制方法.该方法采用全局快速终端滑模面,保证了系统能够从任意初始状态在有限时间内到达滑模面和平衡点.采用模糊神经网络自适应地补偿系统的建模误差和外界干扰,保证了滑模控制在滑模面的运动.文中利用李亚普诺夫稳定性判据推导出了控制器的控制律和模糊神经网络的目标函数,通过模糊神经网络的在线学习.削弱了滑模控制的抖振.仿真结果表明了其有效性.     

基于有限时间的一类时滞非线性切换系统滑模控制

研究一类时滞非线性切换系统的有限时间滑模控制问题.针对所研究的系统模型,构造每个子系统对应的积分滑模面,基于滑模控制理论,设计带有状态时滞的滑模控制器使得每个子系统能在有限时间内到达相应的滑模面上,并对系统中存在的非线性项采用Lipschitz条件进行处理.根据多李亚普诺夫函数、平均驻留时间方法以及分割策略引理,给出滑模趋近段和滑模动态有限时间有界的充分条件,并通过对线性矩阵不等式的求解得到控制器增益.最后,通过一个数值仿真例子验证该设计方法的有效性.     

一类时滞复杂动态网络的鲁棒自适应同步问题

针对一类耦合时滞复杂网络,基于变结构控制研究其同步问题.首先,利用系统的左特征向量函数设计切换面,使网络的运动轨线在期望的同步轨道保持稳定;然后,针对网络中的非线性项已知和未知时,分别设计相应的滑模控制器,使从任意时刻出发的动态网络轨线都能在有限时间内同步到达期望的轨道;最后,利用所设计的控制器对两个算例进行仿真,仿真结果说明了所设计控制器的有效性.     

Finite-Time Synchronization of Complex-Valued Neural Networks with Mixed Delays and Uncertain Perturbations

This paper concerns the problem of finite-time synchronization for a class of complex-valued neural networks (CVNNs) with both time-varying and infinite-time distributed delays (mixed delays). Both the driving and response CVNNs are disturbed by external uncertain perturbations, which may be nonidentical. A simple state-feedback controller is designed such that the response CVNNs can be synchronized with the driving system in a settling time. By using inequality techniques and constructing some new Lyapunov–Krasovskii functionals, several sufficient conditions are derived to ensure the synchronization. It is discovered that the settling time cannot be estimated when the interested CVNNs exhibit infinite-time distributed delays, while it can be explicitly estimated for the CVNNs with bounded delays. The settling time is dependent on both the delays and the initial value of the error system. Finally, numerical simulations demonstrate the effectiveness of the theoretical results.     

Adaptive Backstepping Sliding Mode Control of Tractor-trailer System with Input Delay Based on RBF Neural Network

In this paper, an adaptive sliding mode neural network(NN) control method is investigated for input delay tractor-trailer system with two degrees of freedom. An uncertain camera-object kinematic tracking error model of a tractor car with n trailers with input delay is proposed. Radial basis function neural networks(RBFNNs) are applied to approximate the unknown functions in the error model. A sliding mode surface with variable structure control is designed by using backstepping method. Then, an adaptive NN sliding mode control method is thus obtained by combining Lyapunov-Krasovskii functionals. The controller realizes the global asymptotic trajectories tracking of the kinematics system. The stability of the closed-loop system is strictly proved by the Lyapunov theory. Matlab simulation results demonstrate the feasibility of the proposed method.

     

Sliding mode control of discrete‐time switched systems subject to mode delays

This work considers the sliding mode control problem of a class of discrete‐time uncertain switched systems subject to detecting‐delay on mode signals, which may result in the asynchronous phenomenon between the controller and the switched system. Since the mode information of the controlled system is not available for the controller in time, a mode‐independent sliding surface will be introduced, by which an asynchronous sliding mode controller is designed, whose control gain and robust parameter will be changing according to the controller mode. In the analysis on the stability of the closed‐loop control system and the reachability of the specified sliding surface, the asynchronous characteristics are detailedly investigated. It is shown that the Lyapunov function may be not always decreasing along the state trajectories during the unmatched interval of controller modes and system modes. Nevertheless, it is proven that the state trajectories can be driven into a sliding region around the specified sliding surface in finite time. Finally, some numerical simulation results are provided.     

Finite-Time Synchronization of Memristive Neural Networks with Proportional Delay

The problem of finite-time synchronization for memristive neural networks (MNNs) with proportional delay is considered. Since proportional delay is unbounded and different from infinite-time distributed delay, the classical finite-time analytical techniques are not applicable anymore. First, a discontinuous state feedback controller is designed such that the delayed MNNs achieve drive-response synchronization in a finite settling time. By using Filippov solution and Lyapunov functional method, sufficient conditions are derived. It is shown that, though the proportional delay is unbounded, complete synchronization can still be realized and the settling time can be explicitly estimated. Second, a special adaptive controller is designed for the finite-time problem in order to reduce the control gains. Finally, numerical simulations are given to verify the effectiveness of the theoretical results.     

基于神经网络的多机械臂固定时间同步控制

针对动力学模型未知的多机械臂系统,提出了一种基于神经网络的固定时间终端滑模的位置同步控制器。首先结合相邻交叉耦合同步控制策略,设计固定时间终端滑模面与控制器,保证系统的跟踪误差与同步误差在固定时间内收敛,且收敛时间上界与初始状态无关。其次,设计RBF神经网络权值更新律估计系统多机械臂未知非线性动力学模型,该方法无需对系统模型参数的先验知识。利用Lyapunov函数证明系统的固定时间收敛性与稳定性。最后,仿真结果验证了所提方法的有效性。     

混沌系统的有限时间滑模同步控制

采用滑模控制的方法,研究了两个不同的带有不确定性和外部扰动的混沌系统之间的同步问题。基于Lyapunov稳定性理论和有限时间滑模控制方法,设计了终端滑模控制器来实现两个混沌系统的同步。在设计控制器过程中提出了一个新的非奇异的终端滑模面,并证明它能在有限时间内收敛于零平衡点。通过数值仿真验证了所设计的控制器的有效性。     

具有未知干扰的无人机鲁棒滑模飞行控制

研究无人机飞行稳定性控制问题,由于无人机飞行控制系统存在时变外部干扰,飞行过程中升阴比变化激烈,控制稳定性难度较大。利用滑模控制良好的鲁棒能力提出一种神经网络的鲁棒飞行控制方法。因神经网络有良好非线性逼近能力,可对无人机飞行系统中的不确定进行在线逼近,并将神经网络权值误差引入到权值的自适应律中用以改善系统的动态性能。利用神经网络的组合,设计无人机鲁棒滑模飞行控制器。控制器分为两部分,一部分是等效控制器,另一部分是滑模控制器,能有效减小系统的跟踪误差。最后将所设计的鲁棒滑模控制对无人机飞行姿态控制进行仿真。仿真结果表明,新方法能提高无人机的鲁棒飞行控制能力且能实现无人机姿态的精确跟踪和稳定性控制。     

具有未知参数的混沌系统的有限时间滑模同步控制

针对带有未知参数和非线性输入的两个不同的混沌系统之间的同步问题进行研究. 提出一个相比于传统滑模面具有更快收敛速度的终端滑模面, 并结合自适应控制理论和滑模控制理论, 设计一个自适应滑模控制律, 使同步误差在有限时间内收敛到滑模面, 并沿滑模面在有限时间内收敛到零点, 最终实现两个不同的混沌系统之间的同步. 最后, 以带有不确定性和外部扰动的Lorenz 系统和Liu 系统为例进行数值仿真, 仿真结果表明, 同步误差在有限时间内收敛到零点, 从而验证了所设计控制律的有效性和可行性.     

Finite-time output feedback control of uncertain switched systems via sliding mode design

The problem of sliding mode control (SMC) is investigated for a class of uncertain switched systems subject to unmeasurable state and assigned finite (possible short) time constraint. A key issue is how to ensure the finite-time boundedness (FTB) of system state during reaching phase and sliding motion phase. To this end, a state observer is constructed to estimate the unmeasured states. And then, a state estimate-based SMC law is designed such that the state trajectories can be driven onto the specified integral sliding surface during the assigned finite time interval. By means of partitioning strategy, the corresponding FTB over reaching phase and sliding motion phase are guaranteed and the sufficient conditions are derived via average dwell time technique. Finally, an illustrative example is given to illustrate the proposed method.