无刷直流电机调速系统神经网络自适应滑模变结构控制

为了提高无刷直流电机调速驱动系统的性能,提出神经网络自适应滑模变结构控制策略。推导无刷直流电机端电压与转速之间的微分方程,运用滑模变结构控制理论,通过调节端电压来实现转速控制;为了有效抑制系统在滑模切换面上的抖振采用自适应算法调整滑模增益的大小;从实际应用的角度出发,利用神经网络对非线性函数的任意精度拟合性,设计径向基函数神经网络估计器对控制量中广义扰动进行动态估计。仿真和实验结果表明采用本文提出的方法控制无刷直流电机,超调量小,速度响应快,控制精度高,且系统对各种干扰和参数摄振具有较强的鲁棒性,动、静态性能均优于PID控制。     

基于改进型PI趋近律的无刷直流电机滑模控制

针对无刷直流电机模型参数时变的特性,为解决传统PID控制易受模型参数变化影响的弊端与传统滑模控制(SMC)的抖振问题,将分数阶PI与变速趋近律结合,提出了一种改进型PI趋近律的无刷直流电机滑模控制算法。该算法通过对趋近律的改进,既避免了模型参数变化的影响,又减小了系统输出的抖振。仿真与实验结果表明,与传统滑模控制、传统PI控制相比,具有响应快、鲁棒性强、超调小、抖振小等优点,是一种更优的控制方法。     

基于神经网络的PMSM自适应滑模控制

结合滑模控制和神经网络各自的优点,对永磁同步电机(PMSM)提出了一种基于神经网络的PMSM自适应滑模控制方案.首先设计了带积分操作的滑模变结构位置控制器,通过递归神经网络的在线学习来实时估计系统参数变化和外部负载扰动等不确定性的界限,减小滑模控制器的控制量.进而,在滑模控制器中又引入饱和函数取代符号函数,进一步减弱"抖振"现象.理论分析和实验仿真对比研究的结果表明所提出方法具有优越的动态性能和鲁棒性.     

基于新型趋近率的无刷直流电机滑模变结构控制

为提高无刷直流电机控制系统的快速跟踪性能,提出了一种新的滑模变结构速度控制策略.为了有限抑制系统在滑模切换面上的抖振,提高系统鲁棒性,在建立无刷直流电动机数学模型的基础上,提出了一种新型指数趋近率的变结构控制策略,使切换函数快速进入滑模面,有效地提高了系统的性能.仿真实验结果表明,相比传统的PI控制,该系统在快速性、稳定性、准确性方面有显著提高,对参数扰动具有较强的鲁棒性.     

基于变结构控制发动机转速性能研究

在发动机转速性能优化过程中,针对于传统滑模变结构控制在发动机转速控制中出现的转速抖振现象造成系统不稳定。传统的滑模变结构控制方法不能解决高频抖动问题,无法直接应用到系统中。为解决上述问题,提出通过采用模糊滑模变结构(F-SMC)的控制方法,消除高频抖振对控制器性能的影响.首先,在变结构控制变量的基础上引入模糊控制分量;其次对模糊控制分量进行了设计,并在MATLAB软件上进行了计算机仿真.仿真结果表明,采用模糊滑模方法使汽车发动机转速控制器可以提高系统的鲁棒性。     

无刷直流电动机模糊滑模变结构控制的研究

无刷直流电动机的参数强耦合、高度非线性特性增加了对其速度控制的难度,针对上述特点,为提高系统控制的精度和稳定性,设计了一种模糊滑模变结构速度控制器。在传统的滑模变结构控制基础上,引入模糊控制,建立基于等效控制和切换控制的模糊系统,通过隶属度函数的变化来有效地消弱单纯滑模变结构控制所带来的抖振现象,而不牺牲滑模变结构控制对系统参数摄动和外界扰动的强鲁棒性以及快速性等优点。在Matlab软件上仿真结果表明控制器具有良好的控制性能和鲁棒性能。     

基于线性化反馈的径向神经滑模控制器设计

针对非线性系统,采用径向神经网络逼近及自适应控制方法,利用线性化反馈技术,设计一种自适应神经滑模控制器。滑模变结构控制具有独特的鲁棒性能以及对匹配不确定性和外干扰的完全自适应等特点,但容易出现系统抖振问题,将神经网络应用于滑模变结构控制系统的设计中,系统抖振得到抑制。仿真结果也表明将神经网络与滑模控制相结合的方法是行之有效的。     

基于RBF终端滑模观测器的电机转子位置估计

无刷直流电机常采用位置传感器来检测转子位置,这会影响系统的可靠性,增加电机体积和成本。采用无位置传感器控制技术:引入终端滑模面,其具有快速收敛性和良好观测精度,可减少相位滞后问题;采用RBF神经网络来设计观测器的控制策略,将滑模变量作为神经网络输入,输出即为控制策略,简化控制结构。RBF终端滑模观测器将RBF控制与终端滑模控制的优点紧密结合,优化了控制信号,削弱了抖振现象。仿真结果表明,该观测器能快速准确地估计电机的线反电势及电机转速,系统具有良好性能,满足无刷直流电机的工作要求。     

模糊自学习滑模变结构控制的研究及在直线AC伺服系统中的应用

滑模变结构控制（SMSVC）是一种特殊的非线性控制策略,它对参数变化,负载扰动以及系统的不确定性有强的鲁棒性。然而,系统的控制精度和稳定性受到由滑模开关控制所引起的抖振的影响,在本文给出的方法中,抖振可以通过在传统的滑模控制中引入模糊自学习加以有效的抑制,而对SMVSC的强鲁棒性不产生影响,对于直接驱动的AC直线伺服系统的仿真结果显示了这种方法是有效的。     

一种新的二阶滑模控制规律的研究

为了有效地消除传统滑模变结构产生的抖振,提出了一种新的二阶滑模变结构控制方法.该方法将Teminal滑模动态与二阶滑模有效地结合起来,使二阶滑模既具有传统滑模变结构的特点,又能有效地消除传统滑模变结构所产生的抖振.将此方法用于十字梁试验系统,仿真结果表明,该方法确实有效地消除了抖振,而且鲁棒性也很强.     

Chattering suppression methods in sliding mode control systems

The implementation of sliding mode control is often irritated by high frequency oscillations known as “chattering” in system outputs issued by dynamics from actuators and sensors ignored in system modeling. This paper provides analysis of chattering in such systems with unmodeled based on the Lyapunov theory and the describing function method. It also describes various approaches to reduce chattering including methods based on relay control gain adaptation. And for those systems to which the methods are not applicable, chattering frequency control using hysteresis loop will be provided.     

Sliding mode control of a class of underactuated systems

A sliding mode control approach is proposed to stabilize a class of underactuated systems which are in cascaded form. This class of underactuated systems can represent many real applications after transformations. In the controller design, the sliding surface is designed to stabilize the indirectly controlled modes. Its insensitivity to the model errors, parametric uncertainties and other disturbances and its ability to globally stabilize the system are two advantages of the sliding mode controller.     

Adaptive continuous twisting algorithm

Abstract

In this paper, an adaptive continuous twisting algorithm (ACTA) is presented. For double integrator, ACTA produces a continuous control signal ensuring finite time convergence of the states to zero. Moreover, the control signal generated by ACTA compensates the Lipschitz perturbation in finite time, i.e. its value converges to the opposite value of the perturbation. ACTA also keeps its convergence properties, even in the case that the upper bound of the derivative of the perturbation exists, but it is unknown.     

Adaptive sliding mode control and observation

ABSTRACT

This editorial article gives a short introduction to Special Issue of International Journal of Control on Adaptive Sliding Mode Control and Observation.     

A Sufficient Condition for Checking the Attractiveness of a Sliding Manifold in Fractional Order Sliding Mode Control

Stability issues of fractional order sliding mode control laws are analyzed in this paper. For differentiation orders less than unity, it is shown that a stable reaching law in the fractional order case corresponds to a stable reaching law in the integer order case. The contribution of the current study is to explain the stability of the closed loop by the use of the Caputo and Riemann‐Liouville definitions of fractional order differentiation. Copyright © 2011 John Wiley and Sons Asia Pte Ltd and Chinese Automatic Control Society     

Limiting interactions in decentralized control of MIMO systems

Despite the closed-loop performance advantages of centralized multivariable controllers, the great majority of industrial process control applications is still based on decentralized controllers. Because of their single loop structure, decentralized controllers cannot suppress the interactions of the plant, which are only taken into account in the controller tuning phase. Therefore, it would be useful in many cases to delimit in some way the undesirable effects of the coupling between inputs and outputs of the closed loop system. A control algorithm to achieve this goal is developed in the paper. The proposal is applicable even to processes with non-minimum phase transmission zeros (provided their off-diagonal transfer functions have their zeros in the left-half-plane). Based on sliding mode attributes, the proposed scheme can be easily added to a previously designed decentralized control system.     

Quasi-continuous HOSM control for systems with unmatched perturbations

The control of nonlinear systems subject to unmatched perturbations is studied. A new design algorithm is proposed based on the block control and quasi-continuous higher order sliding modes techniques. The proposed method provides for the finite time exact tracking of a smooth desired signal in spite of unmatched perturbations.     

Interpolating sliding mode observer for a ball and beam system

ABSTRACT

A principle of interpolating sliding mode observer is introduced in this paper. The observer incorporates multiple linear observers through interpolation of multiple estimates, which is treated as a type of adaptation. The principle is then applied to the ball and beam system for observation of the slope of the beam from the measurement of the ball position. The linearised model of the ball and beam system using multiple linearisation points is developed. The observer dynamics implemented in Matlab/Simulink Real Time Workshop environment. Experiments conducted on the ball and beam experimental setup demonstrate excellent performance of the designed novel interpolating (adaptive) observer.     

AN ASYMPTOTIC SECOND‐ORDER SMOOTH SLIDING MODE CONTROL

Presented is a method of smooth sliding mode control design to provide for an asymptotic second‐order sliding mode on the selected sliding surface. The control law is a nonlinear dynamic feedback that in absence of unknown disturbances provides for an asymptotic second‐order sliding mode. Application of the second‐order disturbance observer in a combination with the proposed continuous control law practically gives the second‐order sliding accuracy in presence of unknown disturbances and discrete‐time control update. The piecewise constant control feedback is “smooth” in the sense that its derivative numerically taken at sampling rate does not contain high frequency components. A numerical example is presented.