压电陶瓷迟滞神经网络建模与线性补偿控制

针对压电陶瓷的迟滞非线性,本文首先进行实验测量得到压电陶瓷的位移迟滞数据;通过分析实验数据,引入线性方程实现压电陶瓷输入电压与输出位移关系的线性化,并建立了基于多项式拟合算法的神经网络迟滞模型;根据迟滞模型设计前馈控制器,分别采用了前馈开环和前馈结合PID的方法对压电陶瓷迟滞非线性进行补偿控制实验.实验结果表明,采用前馈开环控制,压电陶瓷位移主环迟滞减小了91.84%,位移次环迟滞减小了85.67%,位移跟踪的平均相对误差为2.97%;采用前馈结合PID控制,压电陶瓷位移主环迟滞减小了96.42%,位移次环迟滞减小了88.44%,位移跟踪的平均相对误差为2.04%.证明了该控制方法能有效地抑制压电陶瓷的迟滞非线性.     

基于输入空间扩张的动态迟滞神经网络模型

针对神经网络不能直接用于辨识具有多值映射特征的迟滞非线性的不足, 利用输入空间扩张的方法, 引入动态迟滞算子来反映动态迟滞的速率依赖性, 由迟滞的输入、输入变化率和算子输出构造神经网络的扩张输入空间, 将输出空间的迟滞多值映射转换为在新的扩张输入空间上的一一映射, 从而将神经网络应用到动态迟滞非线性的辨识中. 所建立模型结构简单, 易于实现在线调整. 最后, 使用该方法对压电陶瓷执行器中的动态迟滞进行了辨识.     

微位移系统中压电陶瓷驱动器迟滞建模

针对超精密微位移系统中压电陶瓷驱动器的迟滞非线性问题,提出了一种基于遗传反向传播(BP)神经网络的压电陶瓷迟滞非线性建模方法.通过电涡流位移传感器获取压电陶瓷驱动器不同电压值下所对应的位移值;利用六次多项式拟合获得迟滞的数学模型,从而建立基于遗传BP神经网络的迟滞,模型.实验结果显示:该迟滞模型在神经网络测试下的最大误差为0.082 1 μm,平均绝对误差为0.0158 μm.表明,所建的迟滞模型能够较精确地反映出压电陶瓷驱动器的迟滞特性,同时为微位移控制系统设计提供了一定的理论基础.     

基于RBF神经网络的压电位移台迟滞非线性建模研究

针对压电陶瓷微位移台固有的率相关迟滞非线性特性,以基于play算子的改进PPI模型构建迟滞算子,结合径向基(Radial Basis Function,RBF)神经网络模型,建立描述压电陶瓷微位移台迟滞特性的率相关模型.研究结果表明,在输入信号频率在10 Hz～90 Hz范围内时,模型输出的最大位移误差为0.399 0 μm～0.932 1 μm,均方根误差为0.259 4 μm～0.565 2 μm,相对误差为0.95％～2.48％.验证了基于PI迟滞算子和RBF神经网络的仿真模型能够准确有效的描述压电微位移台的率相关动态迟滞特性,具有较高的频率泛化能力.该方法易于实现,工程适用性强,具有较好的实用价值.     

压电微定位平台神经网络与专家模糊复合控制方法

为了消除压电微定位平台的迟滞非线性特性,实现高精度定位控制,采用具有两个隐含层的BP神经网络建立压电微定位平台的迟滞模型,以精确描述驱动电压与输出位移的迟滞关系;设计一种基于BP神经网络迟滞逆模型的前馈控制器,对迟滞非线性进行补偿,将迟滞非线性近似线性化.为进一步提高定位系统的精度,提出基于迟滞逆模型前馈补偿和专家模糊控制的复合控制方法.仿真结果表明,该复合控制方法可以将压电微定位平台的定位误差控制在0.091μm以内,从而有效地消除迟滞非线性对压电微定位平台定位精度的影响.     

压电陶瓷执行器的动态迟滞非线性特性建模

针对精密定位系统中压电陶瓷执行器的迟滞非线性特性建模问题,提出了一种基于Hammerstein迟滞模型的建模方法。通过引入一个Backlash类的算子来描述迟滞非线性的轮廓。在利用"扩展输入空间法"将迟滞特性的多值映射转换为一一映射的基础上,采用引力搜索算法优化的支持向量回归机建立静态迟滞模型。为体现迟滞的动态特性,用ARX模型表征迟滞环的率相关性,从而建立了Hammerstein级联模型。并从精密定位系统中采集了实测数据,通过电容传感器获取压电陶瓷执行器给定电压下的位移值,对所提出的模型进行了实验。实验表明:该模型具有较好的性能,辨识过程简便且易于工程实现。     

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.     

人工神经网在二维PSD器件非线性修正中的应用

介绍了一种应用人工神经网对二维PSD器件非线性进行修正的方法。对光斑在二维PSD光敏面上的横向位移，以光斑的二维坐标集合为神经网的期望输出，以PSD输出的二维坐标集合为神经网的训练样本，对神经网络进行训练。利用神经网络所具有的非线性映射能力，在训练结束后即可建立PSD输入与输出的近似线性关系。结果表明修正后的PSD器件可以实现任意输入的实时非线性修正。     

基于Preisach模型的深度学习网络迟滞建模

针对传统压电扫描器迟滞模型泛化能力较弱的问题,提出了一种基于Preisach模型的深度学习网络来建立迟滞模型,提高了模型的学习能力和泛化能力.具体而言,首先利用深度学习在深度特征提取方面的优势,建立包含卷积层、池化层、展开层以及深度特征层的深度学习层来提取输入电压信号的特征信息;其次,利用傅里叶变换层计算得到输入信号的频率,并将频率输入到非线性层,构造并输出了与输入信号频率相关的非线性项,该非线性项作为权值函数与Preisach模型的迟滞单元输出相乘,并将乘积叠加得到了频率相关的模型输出向量;最后,将深度学习层输出的特征向量与Preisach模型输出向量点乘,即可得到深度学习网络的最终输出位移.同时利用电容位移传感器采集的16组输入输出信号对深度学习网络进行训练,得到了网络中的权值参数,并利用其他8组输入输出数据对深度网络进行测试,训练和测试结果表明,本文所提出的基于Preisach模型的深度学习网络在得到高精度迟滞模型的同时,提高了模型的泛化能力.     

基于加速度变化的步态识别方法

为辨识压电作动器在应用过程中的率相关迟滞特性,采用基于最小二乘支持向量机理论对压电作动器进行建模,并引入粒子群算法对模型的参数进行优化,模型以当前及历史输入电压、历史输出位移组成的向量作为模型的输入,而以当前的输出位移作为模型的输出进行训练。最后,通过仿真结果和实验结果对比发现,本文建立的压电率相关迟滞模型平均误差为0.0208μｍ,最大误差为0.4290μｍ比Bouc-Wen模型精度高,验证了本文建模方法的可行性。     

Adaptive identification and control of hysteresis in smart materials

Hysteresis hinders the effective use of smart materials in sensors and actuators. This paper addresses recursive identification and adaptive inverse control of hysteresis in smart material actuators, where hysteresis is modeled by a Preisach operator with a piecewise uniform density function. Two classes of identification schemes are proposed and compared, one based on the hysteresis output, the other based on the time-difference of the output. Conditions for parameter convergence are presented in terms of the input to the Preisach operator. An adaptive inverse control scheme is developed by updating the Preisach operator (and thus its inverse) with the output-based identification method. The asymptotic tracking property of this scheme is established, and for periodic reference trajectories, the parameter convergence behavior is characterized. Practical issues in the implementation of the adaptive identification and inverse control methods are also investigated. Simulation and experimental results based on a magnetostrictive actuator are provided to illustrate the proposed approach.     

一种新的磁滞非线性前馈补偿算法

针对超磁致伸缩致动器磁滞非线性特征, 建立了描述其非线性行为的Preisach数学模型, 以F函数法求解了该模型的数值模型. 针对当前致动器非线性前馈补偿控制中迭代和执行效率低的缺点, 将磁滞非线性理解为系统干扰, 提出了一种新的非线性前馈补偿算法, 在求解Preisach逆模型过程中,引入稳态误差信号作为参考变量, 以Sigmoid函数变步长算法进行迭代步长自适应动态调整. 计算机仿真和实验研究均表明,与当前的磁滞模型求逆算法相比, 所提出的算法在保证控制精度的同时可以显著提高系统收敛速度, 大大提高了程序的执行效率.     

Control of hysteresis in smart actuators with application to micro-positioning

Hysteresis in smart material actuators makes the effective use of these actuators quite challenging. The Preisach operator has been widely used to model smart material hysteresis. Motivated by positioning applications of smart actuators, this paper addresses the value inversion problem for a class of discretized Preisach operators, i.e., to find an optimal input trajectory given a desired output value. This problem is solved through optimal state transition of a finite state machine (FSM) that corresponds to the discretized Preisach operator. A state-space reduction scheme for the FSM is developed, which significantly saves the memory and the computation time. Experimental results on micro-positioning control of a magnetostrictive actuator are presented to demonstrate the effectiveness of the proposed approach.     

A hybrid model for rate-dependent hysteresis in piezoelectric actuators

A hybrid model is proposed in this paper to describe the static nonlinear and dynamic characteristics of rate-dependent hysteresis in piezoelectric actuators. In this model, a neural network based submodel is implemented to approximate the static nonlinear characteristic of the hysteresis while a submodel with the first-order differential operators is constructed to describe the dynamic behavior of the hysteresis. In this paper, a special hysteretic operator is proposed to extract the hysteretic feature of the hysteresis. Then, an expanded input space with such special hysteretic operator is constructed. Based on the constructed expanded input space, the neural network can be implemented to approximate the hysteresis phenomenon. The submodel to describe the dynamics is a sum of the weighted first-order differential operators. Finally, the experimental results of applying the proposed method to the modeling of hysteresis in a piezoelectric actuator are also presented.     

Time-delayed dynamic neural network-based model for hysteresis behavior of shape-memory alloys

Shape-memory alloys (SMAs) have received considerable amount of attentions for their engineering applications in recent years. The hysteresis in SMAs is sensitive to the state-varying tendency and frequency. Utilizing past information to estimate the hysteretic behavior gets increasing attention. In this paper, a time-delayed dynamic neural network (TDDNN) is proposed for modeling hysteresis of SMAs in online applications. By introducing a time delay between the input and output response, the TDDNN considers the time delay’s effect on the hysteresis. This proposed network was applied to a SMA wire actuator. Experimental results demonstrate the effectiveness of TDDNN. The identified results obtained by TDDNN are better than those obtained by dynamic neural network without considering the delay information. It demonstrates the importance of introducing the time delay. The different values of time delay item can also affect TDDNN’s identified results.     

一类具有应力相关迟滞非线性智能结构的建模与跟踪控制研究

智能结构在变化的负载下产生的应力相关迟滞非线性严重影响了其在亚微米级跟踪控制中的应用. 因此, 本文提出了一种应力相关Preisach算子用以描述当输入信号与应力同时变化时, 智能结构所产生的迟滞非线性特性. 该应力相关Preisach算子是在经典Preisach算子的基础上, 通过将应力相关项引入密度函数所得到的. 此外, 为了实现逆补偿控制器的设计, 本文提出了应力相关Preisach算子的性质与基于模糊树方法的辨识方案. 继而, 本文设计了一种基于所提算子的前馈逆补偿与反馈控制器结合的复合控制方案, 并将其应用于一类超磁致伸缩智能结构的实时跟踪控制实验中. 实验证明, 与经典Preisach算子相比, 所提出的算子和相应的控制方案能够较好的消除应力相关迟滞非线性的影响并使控制效果显著提高.     

基于流-热耦合模型的GMM智能构件温升控制系统设计

热效应对超磁致伸缩执行器中超磁致伸缩材料性能产生非常大的影响,从而影响超磁致伸缩执行器的定位精度;提出一种简化的强制水冷策略保证磁致伸缩材料温度恒定;同时建立了超磁致伸缩材料智能构件流-热多场耦合的有限元模型,运用COMSOLMultiphysics 3.4软件对模型进行仿真,仿真结果验证了模型的正确,进一步的实验结果证实了提出的温度控制策略的有效性。     

Intelligent modeling and control for nonlinear systems with rate-dependent hysteresis

A new modeling approach for nonlinear systems with rate-dependent hysteresis is proposed. The approach is used for the modeling of the giant magnetostrictive actuator, which has the rate-dependent nonlinear property. The models built are simpler than the existed approaches. Compared with the experiment result, the model built can well describe the hysteresis nonlinear of the actuator for input signals with complex frequency. An adaptive direct inverse control approach is proposed based on the fuzzy tree model and inverse learning and special learning that are used in neural network broadly. In this approach, the inverse model of the plant is identified to be the initial controller firstly. Then, the inverse model is connected with the plant in series and the linear parameters of the controller are adjusted using the least mean square algorithm by on-line manner. The direct inverse control approach based on the fuzzy tree model is applied on the tracing control of the actuator by simulation. The simulation results show the correctness of the approach. Supported by the National Natural Science Foundation of China (Grant No. 60534020), the National Basic Research Program of China (Grant No. G2002cb312205-04), the Research Fund for the Doctoral Program of Higher Education (Grant No. 20070006060), and the Key Subject Foundation of Beijing (Grant Nos. XK100060526, XK100060422)     

超磁致伸缩作动器的率相关Hammerstein模型与H∞鲁棒跟踪控制

利用Hammerstein模型对超磁致伸缩作动器（Giant magnetostrictive actuators,GMA）的率相关迟滞非线性进行建模,分别以改进的 Prandtl-Ishlinskii（Modified Prandtl-Ishlinskii）模型和外因输入自回归模型（Autoregressive model with exogenous input,ARX）代表Hammerstein模型中的静态非线性部分和线性动态部分,并给出了模型的辨识方法. 此模型能在1～100Hz频率范围内较好地描述GMA的率相关迟滞非线性. 提出了带有逆补偿器和H∞鲁棒控制器的二自由度跟踪控制策略,实时跟踪控制实验结果证明了所提策略的有效性.     

Adaptive Consensus Quantized Control for a Class of High-Order Nonlinear Multi-Agent Systems With Input Hysteresis and Full State Constraints

For a class of high-order nonlinear multi-agent systems with input hysteresis, an adaptive consensus output-feedback quantized control scheme with full state constraints is investigated. The major properties of the proposed control scheme are: 1) According to the different hysteresis input characteristics of each agent in the multi-agent system, a hysteresis quantization inverse compensator is designed to eliminate the influence of hysteresis characteristics on the system while ensuring that the quantized signal maintains the desired value. 2) A barrier Lyapunov function is introduced for the first time in the hysteretic multi-agent system. By constructing state constraint control strategy for the hysteretic multi-agent system, it ensures that all the states of the system are always maintained within a predetermined range. 3) The designed adaptive consensus output-feedback quantization control scheme allows the hysteretic system to have unknown parameters and unknown disturbance, and ensures that the input signal transmitted between agents is the quantization value, and the introduced quantizer is implemented under the condition that only its sector bound property is required. The stability analysis has proved that all signals of the closed-loop are semi-globally uniformly bounded. The StarSim hardware-in-the-loop simulation certificates the effectiveness of the proposed adaptive quantized control scheme.