Search

When the piezoceramic actuator (PEA) perform the high-frequency and nanometer precision motion, the hysteresis existing in PEA are not only non-smooth, nonlinear and multi-value mapping, but also rate-depended. It is very difficult to use the traditional identification scheme to model such a system. In this paper, the classical Bouc-Wen model is modified and the dual Bouc-Wen model structure with linearization in parameters is proposed; and then its coefficients are estimated by the Recursive General Identification Algorithm (RGIA). The experimental results are presented to illustrate the performance of the proposed identification scheme and the obtained model can be applied the high-frequency adaptive control.

For compensating the interactive effects and perform the target tracking control in the multi-dimensional piezo-driven stage, a hybrid decoupling control strategy based on the special sandwich model is presented in this paper. In the proposed decoupling control scheme, the coupling effects of each axis in multi-dimensional piezo-driven stages are firstly eliminated by the nonlinear coupling model. Then, the proposed control schemes based on the obtained sandwich model are developed to enable the displacement of each piezo-driven substage to track the modified reference signal. Besides, the robust controller is design to perform the tracking and decoupling control in the piezo-driven. Finally, the control experimental results on the X-Y piezo-driven stages are also presented to verify the proposed hybrid decoupling control scheme.

Because the non-smooth nonlinear dead-zone exists in the precision moving stage, both its tracking accuracy and sensitivity are severely deteriorated. In this paper, a sandwich model with dead-zone is employed to decribe the precision stage. In this case, a two-stage identification scheme is presented to model the sandwich model, which a special signal called the degeneration input is first designed to degenerate the non-smooth nonlinear dead-zone into a static linear function and the Recursive General Identification Algorithm (RGIA) is used to estimate its parameters. Finally, a numerical simulation and the experimental results are presented to illustrate the performance of the proposed identification.

The precision motion stage exists the non-smooth nonlinear dead-zone varying with displacement and saturation simultaneously. In order to accurately describe such an electromechanical motion system, a novel Hammerstein-Wiener system (HWS) is firstly introduced to handle two non-smooth nonlinearities in this paper. Then, a three-step identification scheme based on two separable input signals is presented to estimate the parameters of HWS. Finally, the experimental results obtained by implementing the proposed identification scheme are also presented in this paper.

In this paper, a novel switching non-smooth observer is proposed to deal with the state estimation of compound non-smooth sandwich system with backlash and dead zone. Based on the characteristic of the system, two separate non-smooth state-space functions are constructed to describe the input–output relationships of backlash. Then, a Luenberger-type non-smooth observer is developed based on the two separate non-smooth state-space functions and the backlash and dead zone input output functions. The observer can switch among the different operating zones automatically in terms of the change of real operation conditions. The observability of the system has been analyzed and the convergence of the switching compound non-smooth observer has been verified. Finally motor servo hydraulic systems which can be represented as a compound non-smooth sandwich system have been studied as an application case. The comparisons between the proposed non-smooth scheme and the conventional method are illustrated. It is demonstrated that the proposed non-smooth approach can achieve much better performance in the aspects of state estimation accuracy and state estimation convergence speed than the conventional one.

In this paper, a two-step identification method for a micropositioning stage with piezoelectric actuator is proposed. It is noted that one of the difficulties encountered in identification is that both input and output of the actuator embedded in the stage cannot be measured directly. Moreover, hysteresis existing in piezoelectric actuators is a non-smooth complex nonlinearity. In the proposed modeling method, a sandwich model with hysteresis is used to describe the performance of the micropositioning stage with piezoelectric actuator. In this modeling architecture, the input linear submodel is utilized to describe the behavior of preceded amplifier with filtering circuit, which provides electrical voltage to the piezoactuator, and the output linear submodel is employed to depict the flexural hinge with load, respectively, while a Duhem function embedded in between the input and output linear submodels is employed to describe the hysteresis characteristic of piezoelectric actuator in the stage. At the first step of the identification procedure, a special excitation input is implemented to excite the stage to decompose the hysteresis into a monotonic polynomial within a certain region. Then, the parameters of linear submodels are separated and estimated. Subsequently, at the second step, an input signal that can fully excite the system within the operation region is implemented to excite the stage. Based on the previously estimated linear submodels, both input and output of the piezoactuator are estimated. Then, in terms of the estimated input and output of the piezoactuator, the parameters of the hysteresis submodel are estimated. Finally, experimental results are presented to verify the proposed method.

In order to accurately describe the high-frequency motion characteristics of the Piezoelectric Actuators (PEA), an online modeling method is proposed in this paper. In this identification scheme, a nonlinear discrete hysteretic operator is firstly introduced to handle the non-smoothness and multi-value mapping of the hysteresis. Then, the key term separation technique is employed to obtain a special form of the sandwich model, which its expression is linearized in parameters. Finally, a Recursive General Identification Algorithm (RGIA) is applied to the estimation of the parameters of the sandwich model. The experimental results obtained by implementing the proposed identification scheme are also presented in this paper. It can satisfy the design requirement of the adaptive control the piezoelectric ultra-precision motion systems.

In this paper, a novel non-smooth observer is proposed to handle the soft measurement of the states of the sandwich system with dead zone. The convergence theorem of the non-smooth observer and its proof are given, respectively. After that, the experimental results are presented. The comparison between the proposed non-smooth scheme and the conventional method is illustrated. Finally, the robust observer is proposed for the sandwich system with dead zone and the states’ estimation results of the experimental equipment by the robust observer and the conventional non-robust observer have been compared in order to illustrate the necessity of proposing the robust one.

In this paper, a modeling method of XY micropositioning stage with piezoelectric actuators is proposed. In the modeling scheme, a sandwich model consists of both input and output linear submodels, and an embedded neural-network-based hysteresis submodel is used to describe the motion behavior of each axis of the stage. Moreover, a neural-network-based submodel is constructed to describe the nonlinear interactive dynamics caused by the movement of another axis. Then, a tracking control scheme combined with a nonlinear decoupling control is proposed to compensate for the effect of the interactions between axes and track the reference trajectory. Then, the robust design method for the tracking and decoupling control is discussed. Finally, the experimental results on an XY micropositioning stage are presented.

This paper presents a method of the identification for the rate-dependent hysteresis in the piezoelectric actuator (PEA) by use of neural networks. In this method, a special hysteretic operator is constructed from the Prandtl–Ishlinskii (PI) model to extract the changing tendency of the static hysteresis. Then, an expanded input space is constructed by introducing the proposed hysteretic operator to transform the multi-valued mapping of the hysteresis into a one-to-one mapping. Thus, a feedforward neural network is applied to the approximation of the rate-independent hysteresis on the constructed expanded input space. Moreover, in order to describe the rate-dependent performance of the hysteresis, a special hybrid model, which is constructed by a linear auto-regressive exogenous input (ARX) sub-model preceded with the previously obtained neural network based rate-independent hysteresis sub-model, is proposed. For the compensation of the effect of the hysteresis in PEA, the PID feedback controller with a feedforward hysteresis compensator is developed for the tracking control of the PEA. Thus, a corresponding inverse model based on the proposed modeling method is developed for the feedforward hysteresis compensator. Finally, both simulations and experimental results on piezoelectric actuator are presented to verify the effectiveness of the proposed approach for the rate-dependent hysteresis.

In this paper, an identification scheme for sandwich systems with a dead zone is proposed. In this scheme, a so-called combinational input is proposed to transform the sandwich model with a dead zone into a model with separated parameters. Then, the recursive general identification algorithm can be applied to the parameter-estimation, since the transformed model has a form that is linearized in coefficients. Finally, the proposed method is applied to the identification of an X–Y moving positioning stage.

A method for the identification of the rate-dependent hysteresis in piezoceramic actuators is proposed. In this approach, both a so-called generalized gradient of the output with respect to the input of the hysteresis and the derivative of the input that represents the frequency change of the input are introduced into the input space. Then an expanded input space is established. Thus, the multi-valued mapping of the rate-dependent hysteresis can be transformed into a one-to-one mapping based on the expanded of the input space. In this case, the neural network method can be applied to the modeling of the rate-dependent hysteresis. Finally, the experimental results are presented to illustrate the performance of the proposed approach.

A feedforward compensator for rate-dependent hysteresis in piezoelectric actuators (PEA) is proposed. In this method, a model with parallel structure is proposed to approximate both hysteretic behavior and rate-dependent feature of the PEA. In the model, the rate-independent hysteretic performance is approximated by a neural network based on the so called expanded input space method, while the rate-dependent behavior of the hysteresis is approximated by a network with the sum of weighted first-order difference operators. To transform the multi-valued mapping of the hysteresis into a one-to-one mapping, the expanded input space with a hysteretic operator is constructed. Then, the corresponding neural inverse sub-model is obtained. After that, a compensator is constructed to compensate for the effect of the rate-dependent hysteresis in the PEA. Finally, the experimental results of applying the proposed method to a piezoelectric actuator are presented.

A non-linear robust control strategy for piezoceramic actuator in a micro-positioning stage is proposed in this paper. In order to describe the relations among the amplifier-filter circuit, piezoceramic actuator and load, a sandwich model with hysteresis is developed. For the compensation of hysteresis effect in piezoceramic actuator, the corresponding inverse model of hysteresis is constructed. Then, a non-linear robust control strategy based on the sandwich model with hysteresis is proposed. The experimental results show that the proposed control scheme has obtained promising results.