移动机器人自适应模糊神经滑模控制

针对含有驱动器动力学的非完整移动机器人轨迹跟踪控制问题,以驱动器电压作为控制输入,提出了一种新型非完整移动机器人自适应模糊神经滑模控制算法,采用自适应动态递归模糊神经网络在线估计由于参数不确定和外加干扰而引起的未知时变函数,减小了不确定性的估计误差;结合自适应鲁棒控制器,不但克服了移动机器人所受的参数与非参数的多不确定性问题,同时也确保了机器人对指定轨迹的跟踪;基于Lyapunov方法进行了参数自适应律设计和系统稳定性分析,保证了控制系统跟踪误差的收敛性,仿真结果表明了该方法的有效性。     

基于 actor-critic 框架的在线积分强化学习算法研究

针对轮式移动机器人动力学系统难以实现无模型的最优跟踪控制问题,提出了一种基于actor-critic框架的在线积分强化学习控制算法。首先,构建RBF评价神经网络并基于近似贝尔曼误差设计该网络的权值更新律,以拟合二次型跟踪控制性能指标函数。其次,构建RBF行为神经网络并以最小化性能指标函数为目标设计权值更新律,补偿动力学系统中的未知项。最后,通过Lyapunov理论证明了所提出的积分强化学习控制算法可以使得价值函数,行为神经网络权值误差与评价神经网络权值误差一致最终有界。仿真和实验结果表明,该算法不仅可以实现对恒定速度以及时变速度的跟踪,还可以在嵌入式平台上进行实现。     

神经网络和模糊算法相结合的永磁同步电机的鲁棒控制

提出一种神经网络和模糊理论相结合的控制算法,用于永磁同步电机的控制.该算法用基于BP神经网络的PID算法作为速度控制器,实现控制系统的在线自适应调整;同时用模糊理论算法作为神经网络控制器输出的限制,实现了良好的控制动态性能.在与传统的PI控制仿真比较中,该算法显示出了较好的控制性能,对负载和电机参数的变化不再敏感,且控制器可以在误差较大的时候快速跟踪指令,而在误差较小的时候实现稳定运行.     

一种基于FNNC-PID的水轮机调速系统控制方法

水轮机调速系统是典型的具有非最小相位、非线性,时变特性的复杂控制系统,难以建立精确的数学模型.针对水轮机调节系统的特性,运用模糊控制和神经网络控制的理论,设计了一种模糊神经网络控制器,利用神经网络结构来实现模糊逻辑推理,通过神经网络的学习来优化模糊控制的隶属度函数以及模糊规则.针对模糊控制存在稳态误差的问题,提出了一种FNNC-PID复合控制器,仿真结果表明,这种控制方案可以消除稳态误差,并使系统具有良好的动态性能和鲁棒性,其控制效果优于常规的PID控制.     

扭矩输入有界的机器人模糊PD轨迹跟踪

针对轨迹跟踪控制中机器人关节驱动器输出扭矩受限的问题,提出一种基于模糊自适应PD的输入有界轨迹跟踪控制算法。不同于以往的控制策略,该算法在控制律中引入具有饱和特性的改进反正切函数,以确保扭矩控制输入的有界性,并结合模糊自适应原理实现PD增益的在线自整定,以改善系统的动态特性。通过对位置跟踪误差进行线性滤波得到速度跟踪误差替代信号,使得整个系统的闭环控制仅需位置输出反馈。利用奇异摄动理论对系统进行了稳定性分析,证明在一定约束下的PD增益自整定过程中,仍能保证系统稳定。仿真和比较结果表明,该算法能够在严格保证控制输入有界的前提下,减小超调量,缩短系统调整时间,具有更优的轨迹跟踪性能。     

神经网络和模糊算法相结合的永磁同步电机的鲁棒控制

提出一种神经网络和模糊理论相结合的控制算法，用于永磁同步电机的控制。该算法用基于BP神经网络的PID算法作为速度控制器，实现控制系统的在线自适应调整；同时用模糊理论算法作为神经网络控制器输出的限制，实现了良好的控制动态性能。在与传统的PI控制仿真比较中，该算法显示出了较好的控制性能，对负载和电机参数的变化不再敏感，且控制器可以在误差较大的时候快速跟踪指令，而在误差较小的时候实现稳定运行。     

基于模糊神经网络再励学习控制器设计及其在倒车模型中的应用

考虑到神经网络学习算法的特点，给出了一种基于再励学习的自组织模糊CPN。它结合了模糊自组织CPN和再励算法的优点，在控制过程中在线调整网络结构以及对网络参数学习，学习效率高，控制结构简单。可以不要求受控对象的学习模型，实现在线控制，应用在倒车模型中仿真结果展示了所设计系统的良好控制性能。     

基于模糊神经网络的永磁同步电机控制

针对永磁同步电机控制系统的非线性与电机参数时变易受扰动的特性,提出一种基于RBF神经网络在线自学习模糊自适应控制器,利用模糊推理机产生的分目标学习误差进行RBF神经网络的在线训练,有效地提高了控制系统的品质。算法简单易于实现;仿真证明,系统具有较好的动态性能。     

一类模型不确定非线性系统的反演预测控制

针对一类模型不确定的单输入单输出仿射非线性系统,设计一种使得闭环系统稳定且滚动时域性能指标在线最小化的预测控制器。运用反演(Backstepping)设计思想获得具有待定参数的控制器表达式,其误差导数中的未知函数采用模糊逻辑系统来逼近,通过直接估计模糊系统最优参数向量的范数上界来设计控制器和自适应律,大大降低了在线计算量。理论证明该方法设计的控制器保证了闭环系统所有信号是半全局有界的,并且跟踪误差收敛于零的某一邻域。仿真算例验证了所提出算法的有效性。     

三维成型系统的多轴电机协调优化控制

针对三维成型系统多轴电机之间协调性能差影响产品成型精度的问题,重点分析三维成型机五轴电机的相互关系,提出一种基于模糊神经网络PID的优化算法。利用模糊神经网络对PID参数的修正,实现电机间同步误差的在线整定,完成三维成型机多轴电机的协调优化控制。仿真结果及对相同的笔筒模型的成型效果对比表明,模糊神经网络PID控制相比于传统的PID控制减小了各电机之间的误差,提高了系统的成型精度,该系统具有更好的鲁棒性。     

Observer‐based adaptive neural control of uncertain MIMO nonlinear systems with unknown control direction

This paper investigates adaptive neural network output feedback control for a class of uncertain multi‐input multi‐output (MIMO) nonlinear systems with an unknown sign of control gain matrix. Because the system states are not required to be available for measurement, an observer is designed to estimate the system states. In order to deal with the unknown sign of control gain matrix, the Nussbaum‐type function is utilized. By using neural network, we approximated the unknown nonlinear functions and perfectly avoided the controller singularity problem. The stability of the closed‐loop system is analyzed by using Lyapunov method. Theoretical results are illustrated through a simulation example. Copyright © 2012 John Wiley & Sons, Ltd.     

Reference Model with an Adaptive Hermite Fuzzy Neural Network Controller for Tracking a Synchronous Reluctance Motor

This article presents a reference adaptive Hermite fuzzy neural network controller for a synchronous reluctance motor. Although synchronous reluctance motors are mathematically and structurally simple, they perform poorly under dynamic modes of operation because certain parameters, such as the external load and non-linear friction, are difficult to control. The proposed adaptive Hermite fuzzy neural network controller overcomes this problem, as using the Hermite function instead of the conventional Gaussian function shortens the training time. Furthermore, the proposed adaptive Hermite fuzzy neural network controller uses an online self-tuning fuzzy neural network to estimate the system's lumped uncertainty. The estimation method involves a fuzzy controller with expert knowledge of the initial weight of the neural network. Finally, the Lyapunov stability theory and adaptive update law were applied to guarantee system convergence. In this article, the responsiveness of the adaptive Hermite fuzzy neural network controller and an adaptive reference sliding-mode controller is compared. The experimental results show that the adaptive Hermite fuzzy neural network controller markedly improved the system's lumped uncertainty and external load response.     

Neural network–based direct adaptive robust control of unknown MIMO nonlinear systems using state observer

This paper focuses on the problem of adaptive robust tracking control for a class of uncertain multiple-input and multiple-output (MIMO) nonlinear system. Unlike most previous research studies, model dynamics, disturbances, and state variables are unknown in this paper. A novel observer-based direct adaptive neuro-sliding mode control approach is proposed of which the only required knowledge is the system output. By incorporating the Adaptive Linear Neuron (ADALINE) neural network (NN) into the conventional sliding mode observer, the proposed observer has favorable performance. In the controller, a radial basis function (RBF) NN is constructed to approximate the unknown equivalent control laws and the estimation of the sliding surface is applied as the input. A gain-adaptation sliding mode term is designed to enhance the robustness of the control system. Besides, the free parameters of the ADALINE NN and the RBFNN are updated online by adaptive laws to obtain optimal approximation performance. Finally, the comparative simulations are given to show the effectiveness and merits of proposed scheme.     

Decentralized adaptive control of nonlinear large‐scale pure‐feedback interconnected systems with time‐varying delays

In this paper, an adaptive decentralized neural control problem is addressed for a class of pure‐feedback interconnected system with unknown time‐varying delays in outputs interconnections. By taking advantage of implicit function theorem and the mean‐value theorem, the difficulty from the pure‐feedback form is overcome. Under a wild assumption that the nonlinear interconnections are assumed to be bounded by unknown nonlinear functions with outputs, the difficulties from unknown interconnections are dealt with, by introducing continuous packaged functions and hyperbolic tangent functions, and the time‐varying delays in interconnections are compensated by Lyapunov–Krasovskii functional. Radial basis function neural network is used to approximate the unknown nonlinearities. Dynamic surface control is successfully extended to eliminate ‘the explosion of complexity’ problem in backstepping procedure. To reduce the computational burden, minimal learning parameters technique is successfully incorporated into this novel control design. A delay‐independent decentralized control scheme is proposed. With the adaptive neural decentralized control, only one estimated parameter need to be updated online for each subsystem. Therefore, the controller is more simplified than the existing results. Also, semiglobal uniform ultimate boundedness of all of the signals in the closed‐loop system is guaranteed. Finally, simulation studies are given to demonstrate the effectiveness of the proposed design scheme. Copyright © 2013 John Wiley & Sons, Ltd.     

Adaptive neural network asymptotic tracking control for a class of stochastic nonlinear systems with unknown control gains and full state constraints

This article addresses the issue of adaptive intelligent asymptotic tracking control for a class of stochastic nonlinear systems with unknown control gains and full state constraints. Unlike the existing systems in the literature in which the prior knowledge of the control gains is available for the controller design, the salient feature of our considered system is that the control gains are allowed to be unknown but have a positive sign. By introducing an auxiliary virtual controller and employing the new properties of Numbness functions, the major technique difficulty arising from the unknown control gains is overcome. At the same time, the -type barrier Lyapunov functions are introduced to prevent the violation of the state constraints. What's more, neural networks' universal online approximation ability and gain suppression inequality technology are combined in the frame of adaptive backstepping design, so that a new control method is proposed, which cannot only realize the asymptotic tracking control in probability, but also meet the requirement of the full state constraints imposed on the system. In the end, the simulation results for a practical example demonstrate the effectiveness of the proposed control method.     

Fixed-time adaptive neural tracking control for a class of uncertain multi-input and multi-output nonlinear pure-feedback systems

This study examines the fixed-time adaptive neural network tracking control problem for a class of unknown multi-input and multi-output (MIMO) nonlinear pure-feedback systems. The introduction of the radial basis function resolves uncertain problems of unknown MIMO systems. The mean value theorem is introduced to overcome the controller design problem attributed to the nonaffine structure in pure-feedback systems. Moreover, a novel fixed-time virtual controller and an actual controller are designed to solve the issue of previous single-input and single-output and MIMO systems that have no solution in the negative domain and at the origin in finite- and fixed-time controls. Furthermore, a design method is proposed. The final designed controller ensures that all signals in the system are bounded. Simulation experiments show that the designed fixed-time controller facilitates smaller tracking error of the system compared with other finite- or fixed-time controllers. Furthermore, the selection of appropriate design parameters allows the tracking error to converge on a small neighborhood of the origin in a fixed time.     

Neural network‐based adaptive attitude tracking control for flexible spacecraft with unknown high‐frequency gain

Adaptive control design using neural networks (a) is investigated for attitude tracking and vibration stabilization of a flexible spacecraft, which is operated at highly nonlinear dynamic regimes. The spacecraft considered consists of a rigid body and two flexible appendages, and it is assumed that the system parameters are unknown and the truncated model of the spacecraft has finite but arbitrary dimension as well, for the purpose of design. Based on this nonlinear model, the derivation of an adaptive control law using neural networks (NNs) is treated, when the dynamics of unstructured and state‐dependent nonlinear function are completely unknown. A radial basis function network that is used here for synthesizing the controller and adaptive mechanisms is derived for adjusting the parameters of the network and estimating the unknown parameters. In this derivation, the Nussbaum gain technique is also employed to relax the sign assumption for the high‐frequency gain for the neural adaptive control. Moreover, systematic design procedure is developed for the synthesis of adaptive NN tracking control with L₂ ‐gain performance. The resulting closed‐loop system is proven to be globally stable by Lyapunov's theory and the effect of the external disturbances and elastic vibrations on the tracking error can be attenuated to the prescribed level by appropriately choosing the design parameters. Numerical simulations are performed to show that attitude tracking control and vibration suppression are accomplished in spite of the presence of disturbance torque/parameter uncertainty. Copyright © 2009 John Wiley & Sons, Ltd.     

基于动态面控制的间接自适应神经网络块控制

针对一类可转化为"标准块控制形"的多输入多输出的非线性系统,基于动态面控制技术,提出一种间接自适应神经网络控制器的设计方案.该方法通过引入1阶滤波器,消除了后推设计中由于反复对虚拟控制的求导而导致的复杂性问题,同时完全避免了反馈线性化方法中可能出现的控制器奇异性问题,且无需控制增益矩阵正定、可逆的条件.利用李亚普诺夫方法,证明了闭环系统是半全局一致终结有界,通过适当选取设计常数,跟踪误差可收敛到原点的一个小邻域内.仿真结果表明所提控制方法的有效性.     

神经网络自学习模糊控制及其应用

本文提出了一种新的自学习模糊控制算法。利用改进的ｐｉ－ｓｉｇｍａ神经网络，对模糊控制器的结论参数进行辨识，并不断修正隶属函数，实现了模糊规则的自动更新。这种方法被用于机器人解耦控制，取得了满意的仿真结果。     

基于D-FNN的电网混沌振荡系统的自适应控制

面向电网混沌振荡系统产生的混沌现象,采用D-FNN对电网混沌振荡系统进行网络建模。在动态模糊神经网络(D-FNN)建模的基础上,提出对连续混沌系统的在线自适应控制方法。该控制方法通过D-FNN对连续电网混沌振荡系统进行在线的同步自适应控制,控制效果很好,并通过对具体的电网混沌振荡系统的数值实验证实了该方法对电网混沌振荡系统混沌控制的有效性和可行性。