Robust adaptive neural network control with supervisory controller

The problem of direct adaptive neural network control for a class of uncertain nonlinear systems with unknown constant control gain is studied in this paper. Based on the supervisory control strategy and the approximation capability of multilayer neural networks (MNNs), a novel design scheme of direct adaptive neural network controller is proposed.The adaptive law of the adjustable parameter vector and the matrix of weights in the neural networks and the gain of sliding mode control term to adaptively compensate for the residual and the approximation error of MNNs is determined by using a Lyapunov method. The approach does not require the optimal approximation error to be square-integrable or the supremum of the optimal approximation error to be known. By theoretical analysis, the closed-loop control system is proven to be globally stable in the sense that all signals involved are bounded, with tracking error converging to zero.Simulation results demonstrate the effectiveness of the approach.     

Decentralized model reference adaptive sliding mode control based on fuzzy model

1.INTRODUCTIONIn recent years,decentralized control of interconnect-ed systems has become an i mportant and challengingtopic[1~3].However,it is usually difficult to modelthe systemexactly due to the complex real environ-ment.In the past several years,active research hasbeen carried out in controller design based on univer-sal approxi mators,such as fuzzy control and neuralnetwork control[4~6].For a class of SISO nonlinearsystems,adaptive fuzzy control approaches were pro-posed based on …     

Decentralized adaptive fuzzy control of time-delayed interconnected systems with unknown backlash-like hysteresis

The problem of decentralized adaptive fuzzy control for a class of time-delayed interconnected nonlinear systems with unknown backlash-like hystersis is discussed. On the basis of the principle of variable structure control （VSC） and by using the fuzzy systems with linear adjustable parameters that are used to approximate plant unknown functions, a novel decentralized adaptive fuzzy control strategy with a supervisory controller is developed. A general method, which is modeled the backlash-like hysteresis, is proposed and removes the assumption that the boundedness of disturbance, and the slope of the backlash-like hystersis are known constants. Furthermore, the interconnection term is supposed to be pth-order polynomial in time-delayed states. In addition, the plant dynamic uncertainty and modeling errors are adaptively compensated by adjusting the parameters and gains on-line for each subsystems. By theoretical analysis, it is shown that the closed-loop fuzzy control systems are globally stable, with tracking error converging to zero. Simulation results demonstrate the effectiveness of the approach.     

一类切换系统的降阶输出反馈滑模控制

研究了一类含有非线性扰动切换系统的鲁棒镇定问题.基于滑模控制技术,提出了降阶输出反馈鲁棒控制策略,给出了系统的滑动模态可达条件,同时设计了该条件下各子系统的输出反馈滑模控制器,并根据李雅普诺夫函数和状态增益给出了离散切换策略,确保闭环切换系统是渐近稳定的.通过数值仿真验证了设计方法的有效性.     

Neural network based adaptive sliding mode control of uncertain nonlinear systems

The purpose of this paper is the design of neural network-based adaptive sliding mode controller for uncertain unknown nonlinear systems. A special architecture adaptive neural network, with hyperbolic tangent activation functions, is used to emulate the equivalent and switching control terms of the classic sliding mode control (SMC). Lyapunov stability theory is used to guarantee a uniform ultimate boundedness property for the tracking error, as well as of all other signals in the closed loop. In addition to keeping the stability and robustness properties of the SMC, the neural network-based adaptive sliding mode controller exhibits perfect rejection of faults arising during the system operating. Simulation studies are used to illustrate and clarify the theoretical results.     

基于观测器的不确定非线性系统L1自适应控制

针对一类严格块反馈型不确定非线性系统,采用反步设计方法提出了一种新的基于滑模状态观测器的L1自适应控制方案。由于系统状态不完全可测,首先设计滑模观测器对系统状态进行估计,并分析观测器的收敛性,在此基础上,通过反步法得到系统的理想控制律。为了消除反步控制中的“项数膨胀”,引入非线性跟踪微分器来提取理想控制律的微分信号。为提高系统响应的瞬态性能,消除控制输入的高频振荡,引入L1自适应控制对控制律进行改进,通过理论推导证明了闭环系统的稳定性。最后通过数值仿真算例验证了所设计控制方案具有快速的收敛性、良好的跟踪性等期望性能。     

一类非线性系统的自适应模糊滑模定位控制

针对一类由多子系统组成的,具有建模误差和未知不确定性的多变量非线性系统,提出了一种自适应鲁棒定位控制方案。分别在系统数学模型已知或未知的情形下,通过对不确定性的未知范数界描述,基于Lyapunov理论和Barbalat引理,给出了滑模鲁棒控制器的综合设计方法及其自适应控制律,保证整个闭环误差系统的稳定。该方法减少了对系统模型精确度的依赖,避免了传统方法对不确定性的人为预估行为。最后,通过船舶动力定位系统的控制仿真表明了本方法的有效性。     

基于观测器的非线性系统直接自适应模糊控制

针对状态不可测的一类不确定非线性系统,利用模糊系统的逼近能力,提出一种基于观测器的直接自适应模糊控制方案。该方案通过引入观测量的调制函数,使得模糊系统的输入集在有界闭区域上,从而取消了观测量有界的假设。利用Kalman Yacubovitch Popov定理及李亚普诺夫函数,证明了闭环控制系统所有信号是有界的,跟踪误差收敛到零。仿真结果表明了该方法的有效性。     

新型神经网络模型参考自适应控制系统设计

针对任意复杂非线性系统，即控制器具有不可分离结构的离散非线性系统，提出新型神经网络模型参考自适应控制。该方案的提出简化了基于神经网络的模型参考自适应控制系统的设计，只需一个神经网络辨识器。统一了任意神经网络模型参考自适应控制的设计方法，更接近于工程实际。仿真结果证明了该方案的合理性和有效性。     

一类纯反馈非仿射非线性系统的自适应神经网络变结构控制

研究了一类非仿射的纯反馈单输入单输出非线性系统。针对此系统,在中值定理、神经网络参数化和解耦Backstepping的基础上,提出了一种自适应变结构神经网络控制策略,而且所给出的定理证明闭环系统的所有信号在平衡点上是半全局一致有界的。通过对一个非仿射CSTR对象的仿真验证了该方法的有效性。     

基于神经网络的机械臂分散自适应跟踪控制

提出了一种适用于机械臂的基于神经网络的分散自适应轨迹跟踪控制方法。将机械臂的轨迹跟踪控制系统考虑为由多个非线性关联组成的不确定性复杂系统,采用分散控制方法进行控制器设计。在对每一子系统设计控制器时,采用直接反馈线性化,利用控制器构建伪线性系统,并引入神经网络自适应环节消除干扰、关联及逼近误差,从而使所提出的控制方法能够保证系统状态有较高的跟踪精度,且算法简单,易于实现。仿真表明,该算法能保证较高的跟踪效果。     

一类不确定离散网络系统的全局滑模控制

针对动态传输控制协议(transmission control protocol, TCP)网络的拥塞问题,基于离散全局滑模控制理论设计了一种主动队列管理(active queue management, AQM)算法,该算法消除了滑模控制的到达阶段,保证网络系统在整个控制过程中的鲁棒性。为了减弱离散指数趋近律的抖振现象,给出了一种改进的滑模趋近律,使得复杂的网络系统具有良好的性能。仿真结果验证了所设计的控制器对活动的TCP连接数、链路带宽及往返时延的不确定性具有很强的稳定性和鲁棒性。     

Indirect adaptive fuzzy control for a class of nonlinear discrete-time systems

An indirect adaptive fuzzy control scheme is developed for a class of nonlinear discrete-time systems. In this method, two fuzzy logic systems are used to approximate the unknown functions, and the parameters of membership functions in fuzzy logic systems are adjusted according to adaptive laws for the purpose of controlling the plant to track a reference trajectory. It is proved that the scheme can not only guarantee the boundedness of the input and output of the closed-loop system, but also make the tracking error converge to a small neighborhood of the origin. Simulation results indicate the effectiveness of this scheme.     

分散自适应模糊滑模控制器与车辆跟随控制

针对一类不确定非线性大系统提出了一种新的组合型自适应分散模糊滑模控制算法.借助模糊逻辑系统与滑模方法设计的控制器加权组合了间接与直接自适应两种模糊控制器,可以同时融合被控对象知识与控制知识两种模糊信息来提高自适应效果.在一类大系统框架下基于统一的格式提出了间接与直接自适应模糊控制设计方案,并进一步消除了现有各种相关控制算法的缺陷.闭环大系统被证明是一致渐近稳定的.该算法应用于自动公路系统车辆的跟随控制,仿真结果表明提出的组合型自适应模糊控制系统比通常间接或直接自适应模糊控制系统具有更好的跟踪性能.     

非线性系统执行器死区故障的鲁棒自适应控制

针对一类具有不对称执行器死区故障的不确定非线性系统,基于反推滑模控制原理提出了一种神经网络鲁棒自适应控制方案。通过简化死区故障模型,取消了模型倾斜度相等和边界对称条件,结合动态面控制避免了传统反推设计方法存在的计算复杂性问题。所提控制方案取消了控制方向已知的条件,消除了执行器死区故障的影响,使得系统输出趋于给定参考轨迹的一个小领域。仿真结果验证了该方法的有效性。     

Adaptive neural network tracking control for a class of unknown nonlinear time-delay systems

1 .INTRODUCTIONNeural network (NN) control has made great pro-gress in past decades[1 ~4]. In Ref .[1] , adaptivebounding design technique was applied to adaptiveneural control for a class of strict-feedback nonlin-ear systems . The requirement of a known boundon the network reconstruction error was removed.By introducing an integral Lyapunov function,anadaptive NN control approach was proposed forunknown strict-feedback nonlinear systems[2],where the controller singularity problem was …     

实现混沌系统同步的自适应变结构观测器方法

针对一类状态不能全部测量且含有不确定项的混沌系统,采用自适应滑模变结构观测器方法,设计了相应的控制律和自适应律,实现了混沌系统的同步控制。所给控制策略,不仅对参数变化和噪声干扰具有很强的鲁棒性,而且不需要已知不确定项的上界,也不需要串联惯性滤波器。理论分析与仿真结果表明了该方法的有效性。     

含非线性输入的非线性系统自适应模糊控制

讨论一类含非线性输入的非线性系统的自适应模糊控制问题。首先运用隐函数存在定理证明系统的理想控制器的存在性,利用I型模糊逻辑系统对该理想控制器进行在线逼近,提出了一种具有监督器的自适应模糊滑模控制器设计的新方案。该方案通过直接自适应模糊控制器与监督控制器的适当切换,保证了闭环系统的稳定性,由此确定出建模的有界区域,而且通过引入最优逼近误差的自适应补偿项,保证跟踪误差收敛到零。仿真结果表明了该方法的有效性。     

Stability analysis of decentralized adaptive backstepping control systems with actuator failures

In this paper, the authors analyze the stability of a class of interconnected systems with subsystem unmodeled dynamics and dynamic interactions employing decentralized adaptive controllers designed by Wen, Zhou, and Wang (2008) in the presence of actuator failures. It will be shown that the global stability of the remaining closed-loop system is still ensured and the outputs are also regulated to zero when some subsystems break down.     

Direct adaptive control for a class of MIMO nonlinear discrete-time systems

This paper considers the problem of adaptive con-trol for a class of multiple input multiple output （MIMO） nonlinear discrete-time systems based on input-output model with unknown interconnections between subsystems. Based on the Taylor ex-pand technology, an equivalent model in affine-like form is derived for the original nonaffine nonlinear system. Then a direct adap-tive neural network （NN） control er is implemented based on the affine-like model. By finding an orthogonal matrix to tune the NN weights, the closed-loop system is proven to be semiglobal y uni-formly ultimately bounded. The σ-modification technique is used to remove the requirement of persistence excitation during the adaptation. The control performance of the closed-loop system is guaranteed by suitably choosing the design parameters.