A New Adaptive Backpropagation Algorithm Based on Lyapunov Stability Theory for Neural Networks

A new adaptive backpropagation (BP) algorithm based on Lyapunov stability theory for neural networks is developed in this paper. It is shown that the candidate of a Lyapunov function V(k) of the tracking error between the output of a neural network and the desired reference signal is chosen first, and the weights of the neural network are then updated, from the output layer to the input layer, in the sense that DeltaV(k)=V(k)-V(k-1)<0. The output tracking error can then asymptotically converge to zero according to Lyapunov stability theory. Unlike gradient-based BP training algorithms, the new Lyapunov adaptive BP algorithm in this paper is not used for searching the global minimum point along the cost-function surface in the weight space, but it is aimed at constructing an energy surface with a single global minimum point through the adaptive adjustment of the weights as the time goes to infinity. Although a neural network may have bounded input disturbances, the effects of the disturbances can be eliminated, and asymptotic error convergence can be obtained. The new Lyapunov adaptive BP algorithm is then applied to the design of an adaptive filter in the simulation example to show the fast error convergence and strong robustness with respect to large bounded input disturbances     

Adaptive reliable guaranteed performance control of uncertain nonlinear systems by using exponent‐dependent barrier Lyapunov function

This paper investigates the issue of adaptive reliable tracking control for a class of uncertain nonlinear parametric strict‐feedback systems under actuator faults. To guarantee better transient performance of adaptive systems especially when actuator faults occur, a novel prescribed performance bounds (PPBs) method based on exponent‐dependent barrier Lyapunov function is developed. Differing from the existing results where the control schemes have introduced the strictly monotone smooth function to achieve constrained error transformation, the proposed PPBs scheme is designed by using the time‐varying barriers to constrain the error trajectories, which accurately characterizes the convergence rates and convergence bounds of errors. Finally, under the framework of backstepping technique and Lyapunov stability theorem, an adaptive reliable controller is designed to ensure that all the closed‐loop signals are semiglobally uniformly ultimately bounded with the tracking errors converging to the specified PPBs. Simulation results demonstrate the effectiveness of the proposed approach.     

An adaptive fuzzy controller based on sliding mode for robotmanipulators

This paper considers adaptive fuzzy control of robotic manipulators based on sliding mode. It is first shown that an adaptive fuzzy system with the system representative point (RP, or as is often termed, a switching function in variable structure control (VSC) theory) and its derivative as inputs, can approximate the robot nonlinear dynamics in the neighborhood of the switching hyperplane. Then a new method for designing an adaptive fuzzy control system based on sliding mode is proposed for the trajectory tracking control of a robot with unknown nonlinear dynamics. The system stability and tracking error convergence are also proved by Lyapunov techniques.     

基于随机牛顿算法的离散系统自适应参数估计

针对一类离散系统,提出一种基于随机牛顿算法的自适应参数估计新框架,相较于已有的参数估计算法,所提出方法仅要求系统满足有限激励条件,而非传统的持续激励条件.所提出算法的核心思想在于通过对原始代价函数的修正,在使用当前时刻误差信息的基础上融入历史误差信息,进而通过对历史信息和历史激励的复用使得持续激励条件转化为有限激励条件;然后,为了解决传统算法收敛速度慢的问题并避免潜在的病态问题,采用随机牛顿算法推导出参数自适应律,并引入含有历史信息的海森矩阵作为时变学习增益,保证参数估计误差指数收敛;最后,基于李雅普诺夫稳定性理论给出不同激励条件下所提出算法的收敛性结论和证明,并通过对比仿真验证所提出算法的有效性和优越性.     

Adaptive Repetitive Control for a Class of Nonlinearly Parametrized Systems

In this note, Lyapunov-based adaptive repetitive control is presented for a class of nonlinearly parametrized systems. Through the use of an integral Lyapunov function, the controller singularity problem is elegantly solved as it avoids the nonlinear parametrization from entering into the adaptive control and repetitive control. Global stability of the adaptive system and asymptotic convergence of the tracking error are established, and tracking error bounds are provided to quantify the control performance. Both partially and fully saturated learning laws are analyzed in detail, and compared analytically.     

Improved convergence of Lyapunov model reference adaptive systems by a parameter misalignment function

A modified approach is presented for the design of Lyapunov model reference adaptive systems. The approach introduces the use of a function of the parameter misalignment, which can be generated with minor demands for additional hardware. Simulation confirms the acceleration in convergence of the system error compared to previous design procedures.     

一种含噪声不确定多输入多输出非线性时变系统的最优跟踪控制

如何在确保实时性能的前提下,将系统耦合、随机因素、时变特性和不确定非线性的影响一同最小化具有重要意义.为此,本文提出了一种基于自适应多维泰勒网(MTN)的优化控制方案,包括MTN控制器(MTNC)和MTN滤波器(MTNF).首先,设计基于强化学习和自适应动量因子的改进梯度法来调节MTNC权值以快速响应被控对象的不确定性和时变特性,实现最优控制;证明闭环系统稳定性.而后,通过Lyapunov稳定性理论设计MTNF权值更新律,使动态误差指数收敛到零;恰当选择Lyapunov函数来构造具有全局最小值的能量空间并对MTNF的Lyapunov特性进行分析;证明MTNF误差的收敛速度和收敛区域,避免奇点问题.最后,仿真结果表明所提出的控制器和滤波器可在较短的时间内获得更高的精度.     

神经网络自适应滑模控制的不确定机器人轨迹跟踪控制

提出一种针对机器人跟踪控制的神经网络自适应滑模控制策略。该控制方案将神经网络的非线性映射能力与滑模变结构和自适应控制相结合。对于机器人中不确定项,通过RBF网络分别进行自适应补偿,并通过滑模变结构控制器和自适应控制器消除逼近误差。同时基于Lyapunov理论保证机器手轨迹跟踪误差渐进收敛于零。仿真结果表明了该方法的优越性和有效性。     

Adaptive tracking control of leader–follower systems with unknown dynamics and partial measurements

In this paper, a decentralized adaptive tracking control is developed for a second-order leader–follower system with unknown dynamics and relative position measurements. Linearly parameterized models are used to describe the unknown dynamics of a self-active leader and all followers. A new distributed system is obtained by using the relative position and velocity measurements as the state variables. By only using the relative position measurements, a dynamic output–feedback tracking control together with decentralized adaptive laws is designed for each follower. At the same time, the stability of the tracking error system and the parameter convergence are analyzed with the help of a common Lyapunov function method. Some simulation results are presented to validate the proposed adaptive tracking control.     

Exponential Continuous Non-Parametric Neural Identifier With Predefined Convergence Velocity

This paper addresses the design of an exponential function-based learning law for artificial neural networks (ANNs) with continuous dynamics. The ANN structure is used to obtain a non-parametric model of systems with uncertainties, which are described by a set of nonlinear ordinary differential equations. Two novel adaptive algorithms with predefined exponential convergence rate adjust the weights of the ANN. The first algorithm includes an adaptive gain depending on the identification error which accelerated the convergence of the weights and promotes a faster convergence between the states of the uncertain system and the trajectories of the neural identifier. The second approach uses a time-dependent sigmoidal gain that forces the convergence of the identification error to an invariant set characterized by an ellipsoid. The generalized volume of this ellipsoid depends on the upper bounds of uncertainties, perturbations and modeling errors. The application of the invariant ellipsoid method yields to obtain an algorithm to reduce the volume of the convergence region for the identification error. Both adaptive algorithms are derived from the application of a non-standard exponential dependent function and an associated controlled Lyapunov function. Numerical examples demonstrate the improvements enforced by the algorithms introduced in this study by comparing the convergence settings concerning classical schemes with non-exponential continuous learning methods. The proposed identifiers overcome the results of the classical identifier achieving a faster convergence to an invariant set of smaller dimensions.      

Indirect adaptive self-organizing RBF neural controller design with a dynamical training approach

This study proposes an indirect adaptive self-organizing RBF neural control (IASRNC) system which is composed of a feedback controller, a neural identifier and a smooth compensator. The neural identifier which contains a self-organizing RBF (SORBF) network with structure and parameter learning is designed to online estimate a system dynamics using the gradient descent method. The SORBF network can add new hidden neurons and prune insignificant hidden neurons online. The smooth compensator is designed to dispel the effect of minimum approximation error introduced by the neural identifier in the Lyapunov stability theorem. In general, how to determine the learning rate of parameter adaptation laws usually requires some trial-and-error tuning procedures. This paper proposes a dynamical learning rate approach based on a discrete-type Lyapunov function to speed up the convergence of tracking error. Finally, the proposed IASRNC system is applied to control two chaotic systems. Simulation results verify that the proposed IASRNC scheme can achieve a favorable tracking performance.     

Adaptive iterative learning control for a class of non-linearly parameterised systems with input saturations

In this paper, an adaptive iterative learning control scheme is proposed for a class of non-linearly parameterised systems with unknown time-varying parameters and input saturations. By incorporating a saturation function, a new iterative learning control mechanism is presented which includes a feedback term and a parameter updating term. Through the use of parameter separation technique, the non-linear parameters are separated from the non-linear function and then a saturated difference updating law is designed in iteration domain by combining the unknown parametric term of the local Lipschitz continuous function and the unknown time-varying gain into an unknown time-varying function. The analysis of convergence is based on a time-weighted Lyapunov–Krasovskii-like composite energy function which consists of time-weighted input, state and parameter estimation information. The proposed learning control mechanism warrants a L²_{[0, T]} convergence of the tracking error sequence along the iteration axis. Simulation results are provided to illustrate the effectiveness of the adaptive iterative learning control scheme.     

符号型自适应模糊控制器

伸缩因子设计是变论域模糊控制的关键, 也是设计的难点. 借助于Lyapunov综合分析方法, 提出一种符号型自适应模糊控制方案, 避免花费精力设计伸缩因子. 方案中使用符号函数替代输入的伸缩运算, 后件调整仍使用积分调节因子. 因此, 本质上它仍是一种论域可变的模糊控制. 相比较变论域模糊控制, 该方案所需规则少, 稳态精度高, 鲁棒性好. Lyapunov稳定性理论保证了跟踪误差的渐近收敛. 最后, 实例证实了方案的可行性.     

Adaptive control for a class of second-order nonlinear systems with unknown input nonlinearities

An adaptive controller is developed for a class of second-order nonlinear dynamic systems with input nonlinearities using artificial neural networks (ANN). The unknown input nonlinearities are continuous and monotone and satisfy a sector constraint. In contrast to conventional Lyapunov-based design techniques, an alternative Lyapunov function, which depends on both system states and control input variable, is used for the development of a control law and a learning algorithm. The proposed adaptive controller guarantees the stability of the closed-loop system and convergence of the output tracking error to an adjustable neighbour of the origin.     

Adaptive fault‐tolerant control for a class of output‐constrained nonlinear systems

In this work, by incorporating a tan‐type barrier Lyapunov function into the Lyapunov function design, we present a novel adaptive fault‐tolerant control (FTC) scheme for a class of output‐constrained multi‐input single‐output nonlinear systems with actuator failures under the perturbation of both parametric and nonparametric system uncertainties. We show that under the proposed adaptive FTC scheme, exponential convergence of the output tracking error into a small set around zero is guaranteed, while the constraint requirement on the system output will not be violated during operation. In the end, two illustrative examples are presented to demonstrate the effectiveness of the proposed FTC scheme. Copyright © 2015 John Wiley & Sons, Ltd.     

An exponentially stable adaptive friction compensator

This note presents a novel adaptive compensation scheme for Coulomb friction in a servocontrol system. An adaptive observer for estimating the unknown Coulomb friction coefficient is also derived on the basis of the Lyapunov technique. In addition, a linearizing control law is developed to compensate for the friction force and obtain the tracking objective. The proposed adaptive compensation guarantees an exponential convergence for state errors and parameter error, and known adaptive schemes guarantee only an asymptotic (or stable) convergence. Simulation results demonstrate the effectiveness of the proposed method for a single-mass servocontrol system     

非线性严格反馈系统自适应非反步输出反馈控制

针对传统反步控制器设计方法存在复杂度爆炸、参数收敛难、控制奇异、需全系统状态已知等问题,提出一种新的可保证参数收敛的未知系统动态辨识和非反步输出反馈自适应控制方法.首先,通过定义新的状态变量和系统等价变换,将严格反馈系统状态反馈控制转化为标准系统的输出反馈控制,进而设计包含高阶微分器的自适应单步控制器,避免反步递推设计的问题;然后,采用两个神经网络对系统集总未知动态进行估计,避免传统控制方法在未知控制增益在线估计过零引发的奇异问题;最后,构造一种新的自适应算法在线更新神经网络权值确保其收敛到真实值,进而实现对未知系统动态的精准辨识.基于Lyapunov定理的分析表明,跟踪误差和估计误差均可收敛到零点附近紧集.基于液压伺服系统模型的对比仿真验证了所提出方法的有效性和优越性.     

Modified fuzzy adaptive controller applied to nonlinear systems modeled under quasi-ARX neural network

In this article, a fuzzy adaptive controller approach is presented for nonlinear systems. The proposed quasi-ARX neural network based on Lyapunov learning algorithm is used to update its weight for prediction model as well as to modify fuzzy adaptive controller. The improving performances of the Lyapunov learning algorithm are stable in the learning process of the controller and able to increase the accuracy of the controller as well as fast convergence of error. The simulations are intended to show the effectiveness of the proposed method.     

On Adaptive Learning Rate That Guarantees Convergence in Feedforward Networks

This paper investigates new learning algorithms (LF I and LF II) based on Lyapunov function for the training of feedforward neural networks. It is observed that such algorithms have interesting parallel with the popular backpropagation (BP) algorithm where the fixed learning rate is replaced by an adaptive learning rate computed using convergence theorem based on Lyapunov stability theory. LF II, a modified version of LF I, has been introduced with an aim to avoid local minima. This modification also helps in improving the convergence speed in some cases. Conditions for achieving global minimum for these kind of algorithms have been studied in detail. The performances of the proposed algorithms are compared with BP algorithm and extended Kalman filtering (EKF) on three bench-mark function approximation problems: XOR, 3-bit parity, and 8-3 encoder. The comparisons are made in terms of number of learning iterations and computational time required for convergence. It is found that the proposed algorithms (LF I and II) are much faster in convergence than other two algorithms to attain same accuracy. Finally, the comparison is made on a complex two-dimensional (2-D) Gabor function and effect of adaptive learning rate for faster convergence is verified. In a nutshell, the investigations made in this paper help us better understand the learning procedure of feedforward neural networks in terms of adaptive learning rate, convergence speed, and local minima.     

基本积分型李亚普诺夫函数的直接自适应神经网络控制

针对一类具有下三角形函数控制增益矩阵的非线性系统,基于滑模控制原理,并利用 多层神经网络的逼近能力,提出了一种直接自适应神经网络控制器设计的新方案.通过引入积 分型李亚普诺夫函数及残差与逼近误差和的上界函数的自适应补偿项,证明了闭环系统是全局 稳定的,跟踪误差收敛到零.