基于非光滑观测器的间隙三明治系统状态估计

在工业领域, 机械传动系统、液压系统等往往含有间隙特性, 这类系统可以用带间隙的三明治系统描述. 本文针对带间隙的三明治系统特点, 构建了一种非光滑观测器以对系统状态进行估计. 首先根据带间隙三明治系统的特点, 采用分离原理, 建立了描述系统特性的非光滑状态空间方程. 据此构造了能够随系统工作区间变化而自动切换的非光滑观测器, 给出了相应的收敛定理及其证明. 最后通过伺服液压系统的例子, 分别比较了非光滑观测器和传统的观测器对状态的跟踪效果, 比较结果表明非光滑观测器对于带间隙三明治系统状态变量估计的准确性要优于传统的观测器.     

基于非光滑观测器的带死区三明治系统状态估计

在控制工程实际中,许多含有死区的系统可以用带死区的三明治系统描述．本文针对带死区的三明治系统特点,构建了一种非光滑观测器以对系统状态进行估计．首先根据带死区三明治系统的特点,由分离原理,建立了描述系统特性的非光滑状态空间方程．据此构造了能够随系统工作区间变化而自动切换的非光滑观测器,给出了相应的收敛定理及其证明．最后通过仿真,分别比较了非光滑观测器和传统的观测器对状态的跟踪效果,比较结果表明非光滑观测器对于带死区三明治系统状态变量估计的准确性要优于传统的观测器．     

间隙三明治系统的改进卡尔曼状态估计

控制工程中许多实际系统都可以描述为带间隙的三明治系统,由于间隙具有非光滑、局部记忆性和多值映射等复杂非线性特性,使得整个三明治系统的内部状态估计工作具有很大挑战性.首先根据间隙三明治系统的特性引入了几个自动切换函数,采用关键项分离原理,建立了随机噪声干扰下间隙三明治系统的非光滑整体伪线性状态空间模型.针对该系统提出了一种非光滑的改进卡尔曼滤波算法以估计系统状态,其工作机制能够随系统当前工作区间的转变而自动切换模式.仿真和实验结果表明,针对含噪声的间隙三明治系统,非光滑的改进卡尔曼滤波算法对系统状态的估计准确度要高于传统卡尔曼滤波算法.     

含有迟滞Hammerstein系统的非光滑卡尔曼状态估计

迟滞特性具有非光滑、多值映射等复杂特性.如果迟滞环节的末端还存在一个线性子系统,导致迟滞的输出信号不可测,使得整个系统的状态估计工作成为很大的难题,常规的估计方法无法直接应用到这类系统中.本文提出一种新的非光滑卡尔曼滤波器,描述了Hammerstein系统的状态空间方程.据此构造了能够随系统工作区间变化而自动切换的非光滑滤波器.最后通过仿真和实验,比较了非光滑卡尔曼滤波器和传统的卡尔曼滤波器的状态估计效果,比较结果表明非光滑卡尔曼滤波器对于带迟滞的Hammerstein系统状态变量的估计的准确性要优于传统的卡尔曼滤波器.     

基于鲁棒观测器的带间隙三明治系统故障预报

工程实践中常见的带间隙的三明治系统的准确故障预报具有重要的现实意义,为此,本文构建了一种新的动态鲁棒观测器对其进行故障预报.首先,通过将非光滑项转化为干扰项的方法,将间隙非光滑三明治系统转化为可用动态鲁棒观测器设计方法设计的系统.其次,采用零点配置和最小化基准区间观测器的范数(H_∞,F/H_,F)指标的方法确定动态鲁棒观测器的增益矩阵.最后,通过仿真,分别比较了基于非光滑鲁棒观测器和基于传统观测器的故障预报效果,比较结果表明:鲁棒观测器能够及时地准确预报传统观测器无法预报的故障,且有效减少了故障的漏报和错报现象.     

动态迟滞Hammerstein系统的改进卡尔曼状态估计

迟滞特性具有非光滑、多值映射等复杂特性.而在实际的工程中,当输入电压变化频率超过一定的范围时,迟滞的特性是随着输入频率的改变发生变化,使得整个系统的状态估计工作更复杂.本文首先提出一种新的描述动态迟滞的方法,进而描述了动态迟滞Hammerstein系统的状态空间方程,根据此系统在传统卡尔曼滤波器的基础上进行改进得到一种新的非光滑卡尔曼滤波器.最后通过仿真和实验,比较了在输入信号变化频率比较大时,用动态迟滞Hammerstein系统来描述压电陶瓷和采用静态迟滞Hammerstein系统来描述压电陶瓷的特性,非光滑卡尔曼滤波器对这两种含有噪声的模型进行滤波,结果表明由于静态迟滞Hammerstein系统的建模不能很好的描述压电陶瓷的特性,模型存在着误差,因此对系统状态估计的结果也没有用动态Hammerstein系统的误差小,从而说明当输入电压频率变化比较大时研究动态的迟滞Hammerstein模型是很有意义的.     

基于迟滞算子的非平滑三明治系统自适应控制

针对一类具有非平滑的迟滞三明治系统, 提出一种基于神经网络的自适应控制方法. 首先利用神经网络做出了前端动态模块的逆系统实现前端动态模块的近似补偿, 这样将迟滞三明治系统转化成一般的迟滞非线性系统. 然后提出一个迟滞算子将迟滞的多映射转化成一一映射, 基于这个迟滞算子设计了神经网络自适应控制器, 通过Lyapunov方法证明了系统的稳定性并推导出神经网络的权值自适应调整律和控制律. 最后通过仿真验证了该方案的有效性.     

压电陶瓷执行器的非光滑三明治模型辨识与内模控制

压电陶瓷执行器中含有非光滑、多值映射、频率依赖的非线性迟滞特性,然而在实际应用中,压电器件的输入输出信号无法直接测量,常规方法难以进行有效的辨识和控制.本文采用三明治模型来精确描述实际对象,并提出一种基于退化激励信号的两步辨识法解决三明治迟滞模型的辨识问题.最后,基于已辨识的三明治模型,设计一个内模控制器,解决压电陶瓷执行器的精密轨迹控制问题.实验结果表明所提出的辨识和控制方案取得了令人满意的结果.     

对迟滞三明治系统基于Duhem算子的自适应控制

对具有迟滞非线性的三明治系统,设计了基于Duhem算子的神经网络自适应控制器.首先对前端动态子系统进行近似补偿;然后用Duhem算子描述所提出的迟滞状态,用神经网络逼近迟滞状态与迟滞输出的关系,实现对迟滞非线性的建模.基于该迟滞模型并采用伪控制技术设计神经网络自适应控制器,通过Lyapunov方法证明了系统的稳定性,并推导出神经网络的权值自适应调整律和控制律.最后通过仿真验证了该方案的有效性.     

基于退化激励信号的含有迟滞的三明治系统辨识

含有迟滞的三明治系统不仅具有非光滑、多值映射等特性, 而且迟滞环节的输入输出信号还是不能直接测量的, 常规方法难以进行有效的辨识. 本文提出了一种基于退化激励信号的两步辨识方案: 第一步, 设计一个特殊的退化激励信号将迟滞环节退化为一条静态曲线, 从而可以将两端的线性动态环节辨识出来, 解决中间信号不可测的问题; 第二步, 利用已辨识的线性模型重构迟滞环节的输入输出信号, 再采用“扩展输入空间法”建立迟滞环节的神经网络模型. 最后, 在压电超精密运动系统的实验结果表明所提出的建模方法取得了令人满意的结果.     

New scheme for discrete observer design with estimation error covariance assignment

This article presents an innovative method for solving an estimation error covariance assignment problem to design an observer for a stochastic linear system. In the proposed method, the covariance assignment problem is converted to the problem of finding an extra noise-like input to the observer. Using appropriate matrix manipulation, the Riccati equation of the estimation error covariance assignment problem, is converted to a new deterministic linear state-space model. Also, the extra noise-like input to the observer is modelled as an input to the new deterministic linear state-space model. Therefore, all the conventional and well-defined control strategies could be applied and there is no need to solve a complicated Riccati equation. Moreover, using the proposed method, a multi-objective estimation error covariance tracking problem would be easily converted to the problem of controlling a standard deterministic linear state-space system. Based on the integral control method, which is applied to the new state-space model, formulations for the proposed covariance feedback law are presented. The control law results in a stable closed-loop covariance system and assigns a pre-specified covariance matrix to the estimation errors.     

Optimal bilinear observers for bilinear state-space models by interaction matrices

This paper formulates optimal bilinear observers for bilinear state-space models. Observers in bilinear form, as opposed to other nonlinear forms, are required to develop an extension of observer/Kalman filter identification for simultaneous identification of a bilinear state-space model and an associated bilinear observer from noisy input–output measurements. The paper establishes the relationship between the bilinear observer gains and the interaction matrices which are used to convert the original bilinear state-space model to a form that simplifies the identification of such a model. Techniques to find the interaction matrices are developed. In the absence of noises, these matrices produce the gains of the fastest converging observer. In the presence of noises, they minimise the state estimation error in the same manner as a standard steady-state Kalman filter. Numerical examples illustrate both the theoretical and computational aspects of the proposed algorithms.     

Velocity observers for linear mechanical systems subject to single non-smooth impacts

In this paper, we consider the problem of the velocity estimation from position measurement for linear mechanical systems (with multiple degrees of freedom) subject to single non-smooth impacts, both elastic and inelastic (i.e., with coefficient of restitution e=1 and e(0,1), respectively). Through a simple example, it is shown that a classical Luenberger observer is not able to reproduce instantaneously the jumps in the mass velocities, since it recovers the error induced by such jumps only asymptotically: an infinite sequence of impacts can prevent the estimation error to asymptotically go to zero. A new observer structure is proposed for linear mechanical systems subject to single unilateral constraints, that guarantees that the corresponding error dynamics are exponentially stable, also in presence of an infinite sequence of non-smooth impacts. The observer that we propose switches at the impact times, that can be recognized by position measurements only. To validate the proposed observer, both simulation and experimental tests have been carried out and are briefly reported, pointing out the drawbacks and advantages of the observer.     

Observability and controllability analysis for sandwich systems with dead-zone

In this paper, an approach to analyze the observability and controllability of sandwich systems with dead-zone is proposed. In this method, a non-smooth state-space function is proposed to describe the sandwich systems with dead-zone which are also non-smooth nonlinear systems. Then, a linearization method based on non-smooth optimization is proposed to derive a linearized state-space function to approximate the non-smooth sandwich systems within a bounded region around the equilibrium points that we are interested in. Afterwards, both observability and controllability matrices are constructed and the methods to analyze the observability as well as controllability of sandwich system with dead-zone are derived. Finally, a numerical example is presented.     

Velocity observers for non-linear mechanical systems subject to non-smooth impacts

This paper is concerned with the estimation of the velocity variables (when the position variables are the measured outputs) for non-linear mechanical systems subject to non-smooth impacts, both elastic and inelastic (i.e., with coefficient of restitution e=1 and e(0,1), respectively). A reduced-order observer is proposed, which guarantees that the corresponding error system, despite the possible presence of an infinite sequence of non-smooth impacts, is locally exponentially stable. An estimate of the basin of attraction is also given.     

Discrete-time track following controller design using a state-space disturbance observer

This paper presents a method to design a discrete-time track following controller using a state-space disturbance observer. To improve sensitivity, an add-on state-space disturbance observer is introduced to a LQG/LTR track following controller, which does not affect the observer and state feedback poles thereby preserving the separation principle. Therefore disturbance observer design is possible to shape the sensitivity without affecting the stability of the LQG/LTR track following controller. The proposed disturbance observer is designed in state-space without disturbance model such as plant's inverse dynamics, periodic signal generator, and Q filter. Simulation and experimental results verify the effectiveness of the proposed design method using a disturbance observer.