基于ESO的动态解耦方法及其应用仿真*

针对多变量系统控制中的耦合问题,提出了一种基于扩张状态观测器（ESO）的动态解耦方法。该方法将系统输入变量间的耦合作用、被控对象参数时变和外界干扰视为一个总的扰动,用ESO估计该总扰动并反馈到控制器进行补偿,从而实现动态解耦;对解耦后的每个子系统,分别设计出了基于误差最小二乘指标的神经元自适应PID（NAPID）控制器。该方法简化了解耦过程,放松了对系统模型的要求,计算量小、鲁棒性强。最后用该法对蒸馏塔进行控制仿真,仿真时使用混沌优化方法对ESO的参数进行了离线优化,并给出了与模糊PID解耦控制方法对比的     

一种基于反推技术和ESO补偿的解耦控制方法

针对一类非线性多入多出系统,设计了一种新的解耦控制器.首先采用反推技术设计每一个子系统的等效输入;然后用扩张状态观测器(ESO)对各个子系统的耦合、外扰和参数时变等不确定作用进行动态估计,并把估计值引入到反推控制器中进行补偿,最终实现解耦和摔制.利用Lyapunov稳定性理论证明了该方法的收敛性;数例仿真表明了该方法可获得较快的响应速度及较好的解耦效果和鲁棒性.     

基于表贴式永磁同步电机模型改进的自抗扰控制器

在使用自抗扰（Active Disturbance Rejection ControI er,ADRC）控制器的表贴式永磁同步电机（Surface Permanent Magnet Synchronous Motor,SPMSM）系统中,电流环依然存在耦合现象,扩张状态观测器（Extended State Observer,ESO）对于变化较大的扰动难以保证对其的估计精度。为了提高负载的动态性能,提出了改进的自抗扰控制器设计方案。基于表贴式永磁同步电机模型,利用最小二乘算法辨识出电流环参数,对电流环模型进行补偿,实现电流环的精确解耦,减小ESO对扰动估计的压力,提高扰动估计精度。仿真结果表明,改进的自抗扰控制器提高了负载的动态性能。     

热连轧板宽板厚的实用自抗扰解耦控制

针对热连轧板宽板厚多变量系统存在强耦合、大时滞和随机不确定等难题,提出了一种线性自抗扰动态解耦方案.考虑到系统的大时滞问题,在常规的降阶扩张状态观测器(ESO)之前,增加了一个纯时滞环节.为了把所设计的实用自抗扰控制(ADRC)与常规PID控制器进行公平比较,各控制器的最佳参数均采用变尺度混沌优化方法得到.仿真结果表明,优化后的ADRC不仅具有较好的解耦性能,而且对模型参数的不确定性和外扰具有较强的鲁棒性和参数适应性.     

基于优化模型补偿自抗扰的伺服控制方法研究

在永磁同步电机(PMSM)自抗扰控制器(ADRC)系统中,扩张状态观测器(ESO)在扰动较多且幅值变化大时难以保证估计精度,而普通的模型补偿自抗扰控制器的性能又受到参数辨识精度和辨识算法复杂度的限制.针对该问题,提出一种伺服系统优化的模型补偿自抗扰控制方法.以交轴电流和实际转速作为线性扩张状态观测器(LESO)输入,采用二阶LESO对系统总扰动进行观测,将此观测值作为补偿模型补偿到速度环ADRC的ESO中,并在控制量的扰动补偿项中去除该补偿模型,实现模型补偿的目的.仿真和实验结果表明,该方法显著降低了ADRC中ESO要估计量的变化幅度,提高了扰动估计精度,同时不需要进行额外的参数辨识,可实时在线获取补偿模型,具有较好的动态特性与抗扰动能力.     

随机系统的多模型直接自适应解耦控制器

针对多变量离散时间随机系统, 提出了一种采用广义最小方差性能指标的多模型直接自适应解耦控制器. 该多模型控制器由多个固定控制器和两个自适应控制器构成. 固定控制器用以覆盖系统参数的可能变化范围, 自适应控制器用以保证系统的稳定性和提高暂态性能. 该多模型控制器利用矩阵的伪交换性和拟Diophantine方程性质, 基于广义最小方差性能指标, 将随机系统辨识算法和最优控制器设计相结合, 直接辨识出控制器的参数, 通过广义最小方差性能指标中加权多项式的选取,不但实现了多变量系统的动态解耦控制, 而且消除了稳态误差、配置了闭环极点. 文末给出了全局收敛性分析. 仿真结果表明该方法明显优于常规自适应控制器.     

立方体机器人自抗扰平衡控制方法

针对立方体机器人动力学模型多变量、强耦合的问题,提出一种基于自抗扰控制的平衡控制器设计方法.引入虚拟控制量,并在控制量与输出向量之间并行地嵌入多个自抗扰控制器,从而实现对多变量系统的解耦控制,将系统的动态耦合和外部扰动视为各自通道上的自抗扰控制器的总扰动,在为期望姿态安排过渡过程基础上,设计扩张状态观测器对总扰动进行估计并实时补偿.综合采用经验试凑法和带宽法对控制器参数进行整定,对自抗扰控制器系统进行稳定控制、姿态跟踪、抗扰性和鲁棒性实验,并与PID控制系统进行定量对比分析.仿真结果表明,所设计的自抗扰控制器不仅能有效实现立方体机器人的平衡控制,而且较PID控制器具有更好的响应速度、控制精度和强鲁棒性.     

扰动作用下的多移动机器人编队模型预测控制

针对多移动机器人编队中的领航者和跟随者同时受系统内部未建模动态和风向、路面平整度等内外扰动影响而带来的无法保持编队队形的问题,提出了基于扩张状态观测器(ESO)的模型预测(MPC)编队控制方法。首先,建立带有扰动项的领航跟随编队模型,然后分别设计ESO对领航者和跟随者所受扰动进行估计,将该编队模型进行线性化离散化处理作为MPC编队控制器的预测模型,在预测输出方程中引入对扰动的估计结果,最后通过滚动优化求解最优控制律,实现对多移动机器人的鲁棒编队控制。仿真实验结果验证了所提出的控制方法的有效性。     

冷连轧机厚度和张力系统的多变量解耦鲁棒控制

针对冷连轧过程中多变量、强耦合、不确定的厚度和张力系统,提出了基于不变性原理解耦的H∞混合灵敏度鲁棒控制策略.建立了厚度和张力系统的动态耦合模型,应用不变性解耦原理实现了对厚度和张力系统的解耦,以减弱系统的耦合影响,并针对系统动态模型的参数摄动以及外部扰动等不确定性,采用H∞混合灵敏度方法设计了鲁棒控制器以保证系统的鲁...     

基于非线性增益递归滑模的船舶轨迹跟踪动态面自适应控制

针对三自由度全驱动船舶存在模型不确定和未知外部环境扰动的情况,设计出一种基于非线性增益递归滑模的船舶轨迹跟踪动态面自适应神经网络控制方法.该方法综合考虑船舶位置和速度误差之间关系设计递归滑模面,引入神经网络对船舶模型不确定部分进行逼近,设计带σ-修正泄露项的自适应律对神经网络逼近误差与外界环境扰动总和的界进行估计,并应用一种非线性增益函数构造动态面控制律,选取李雅普诺夫函数证明了该控制律能够保证轨迹跟踪闭环系统内所有信号的一致最终有界性.最后,基于一艘供给船进行仿真验证,结果表明,船舶轨迹跟踪响应速度快、精度高,所设计控制器对系统模型参数摄动及外界扰动具有较强的鲁棒性.     

Linear extended state observer based sliding mode disturbance decoupling control for nonlinear multivariable systems with uncertainty

In this paper, the sliding mode dynamic disturbance decoupling tracking control method based on the linear extended state observer (LESO) is proposed for a class of square multivariable nonlinear uncertain system. The model plant contains the known linear dynamics, the unknown nonlinear dynamics and the internal and external disturbances, and the various input-output pairs are interacted. The system states are not available for measurement. An improved LESO is developed. The leading feature which is different from the typical ESO lies in that its extended state does not contain the known linear dynamics. The improved LESO can guarantee the error variables to be uniformly ultimately bounded with respect to a ball whose radius is a function of design parameters. So this ball radius can be arbitrarily as small as desired by tuning design parameters. And we give a simple method by which the gain parameters of LESO can be computed easily. This estimation to the total disturbance of the original system is introduced into the sliding mode control design to complete disturbance rejection and decoupling. Rigorous stability analysis shows that the system output can track the desired signal closely. Finally, a class of mass–spring–damper system is taken to make the numerical simulation analysis to illustrate the effectiveness of the proposed control method.     

基于一种新型控制算法的两电机同步控制系统研究

提出一种新的二阶自抗扰控制器的两电机同步系统控制方案,两个二阶ADRC分别控制速度回路和张力回路,实现系统速度和张力的动态解耦。扩张状态观测器将系统模型内扰、外扰和速度张力之间的耦合影响统一视为系统总扰动进行实时观测和补偿。在西门子S7-300PLC实验平台上完成了跟踪特性和解耦特性测试实验,与传统PID控制相比,方波速度响应时间提高了5.3s;稳态误差提高了4.31%;当速度突变时张力几乎没有影响。结果表明二阶ADRC控制器能够实现两电机同步系统中速度和张力的解耦控制。     

结晶器多变量耦合系统的自抗扰控制

主要研究新颖实用非线性自抗扰控制算法,在结晶器多变量耦合系统中的应用.自抗扰控制主要特性是实时估计对象模型摄动和外扰的总和作用量,并在控制信号中补偿掉,实现不确定性强非线性对象的实时动态反馈线性化.结合控制对象,建立了结晶器多变量耦合自抗扰控制系统.数值仿真试验表明自抗扰耦合控制的协调性、自适应跟随性和抗干扰性优于传统的PID解耦控制.     

倒立摆的自抗扰解耦控制方法研究

针对高阶次、多变量、强耦合的倒立摆系统,提出了基于自抗扰解耦控制方法对其进行控制。该方法采用双通道复合控制的方式,突破摆角和小台车相耦合作用的定位控制难点,完成摆角和小台车精确定位控制。采用该方法无需进行精确的输入变量与输出变量配对,只需要对控制作用的耦合矩阵进行静态解耦,分别构建控制通道。对于模型摄动造成的影响和各子对象间的动态耦合作用可由状态观测器（ESO,extended state observer）来估计并反馈到控制器以进行补偿。最终通过调节控制系统的相关参数,达到摆角和小台车位移相耦合作用的定位控制。仿真结果表明：该方法能够完成摆角和小台车位移相耦合作用的定位控制,瞬态性能和稳态性能好,超调量小,具有良好的控制效果。     

基于扩张状态观测器的动态抗饱和补偿器设计方法

针对一类受外界扰动以及执行器饱和影响的不确定非线性系统,提出一种基于扩张状态观测器的动态抗饱和补偿器设计方法.首先通过将系统的不确定项以及外部扰动作为扩张状态,设计线性扩张状态观测器(ESO)对系统的总扰动进行估计;然后,在控制器中引入对扩张状态的估计值,对系统的总扰动进行补偿,设计了动态抗饱和补偿器,将控制器、观测器以及动态抗饱和补偿器的参数求解问题转化为基于LMI不等式组约束的优化问题,确保系统具有尽可能大的收敛域;最后通过数值仿真验证所提出设计方法的有效性.     

Active disturbance rejection control to MIMO nonlinear systems with stochastic uncertainties: approximate decoupling and output-feedback stabilisation

ABSTRACT

In this paper, we apply the active disturbance rejection control, an emerging control technology, to output-feedback stabilisation for a class of uncertain multi-input multi-output nonlinear systems with vast stochastic uncertainties. Two types of extended state observers (ESO) are designed to estimate both unmeasured states and stochastic total disturbance which includes unknown system dynamics, unknown stochastic inverse dynamics, external stochastic disturbance without requiring the statistical characteristics, uncertain nonlinear interactions between subsystems, and uncertainties caused by the deviation of control parameters from their nominal values. The estimations decouple approximately the system after cancelling stochastic total disturbance in the feedback loop. As a result, we are able to design an ESO-based stabilising output-feedback and prove the practical mean square stability for the closed-loop system with constant gain ESO and the asymptotic mean square stability with time-varying gain ESO, respectively. Some numerical simulations are presented to demonstrate the effectiveness of the proposed output-feedback control scheme.     

Extended state observer for uncertain lower triangular nonlinear systems subject to stochastic disturbance

The extended state observer (ESO) is the most important part of an emerging control technology known as active disturbance rejection control to this day, aiming at estimating “total disturbance” from observable measured output. In this paper, we construct a nonlinear ESO for a class of uncertain lower triangular nonlinear systems with stochastic disturbance and show its convergence, where the total disturbance includes internal uncertain nonlinear part and external stochastic disturbance. The numerical experiments are carried out to illustrate effectiveness of the proposed approach.     

Altitude control for flexible wing unmanned aerial vehicle based on active disturbance rejection control and feedforward compensation

For achieving the accurate trajectory tracking of the flexible wing unmanned aerial vehicle in the complicated missions, especially the vertical component, a feedforward compensation unit–based active disturbance rejection control (ADRC) is proposed. In ADRC, the internal dynamics and complicated influence of the total disturbance will be estimated and dynamically compensated by extended state observer (ESO). It puts a very high request on the observation ability of ESO with the unpredictable external disturbance, complex internal coupling influence, and the strong nonlinear characteristic of the proposed system. For this reason, by deeply analyzing the model of this system, the varying attitude influence on the altitude control will be deduced. Then, this influence will be compensated previously by a feedforward compensation unit. Through the previous compensation of the calculable part of the internal dynamics and total disturbance, the burden of ESO can be reduced largely. In this way, it improves the control effect of the ADRC with better observation precision of ESO. After that, based on the hardware‐in‐the‐loop simulation, the effectiveness of the proposed method is verified completely with the complicated flight missions. The robustness of the control effect and observation ability of ESO are also verified by the Monte Carlo simulation. At last, the results of actual flight experiment prove the advancement and practicability of the proposed ADRC method.     

Extended state observer for MIMO nonlinear systems with stochastic uncertainties

ABSTRACT

In this paper, both linear extended state observer (ESO) and nonlinear ESO with homogeneous weighted functions are proposed for a class of multi-input multi-output (MIMO) nonlinear systems composed of coupled subsystems with large stochastic uncertainties. The stochastic uncertainties in each subsystem including internal coupled unmodelled dynamics and external stochastic disturbance without known statistical characteristics are lumped together as the stochastic total disturbance (extended state) of each subsystem. The linear ESO and nonlinear ESO are designed separately for real-time estimation of not only the unmeasured state but also the stochastic total disturbance of each subsystem. The practical mean square convergence of these two classes of ESOs are developed. Some numerical simulations are presented to demonstrate the effectiveness of the ESOs with the advantages of smaller peaking values and more accurate estimation by the nonlinear ESO.