Model-based spatial feedforward for over-actuated motion systems

In high-performance motion systems, e.g. wafer-stages and pick-and-place machines, there is an increasing demand for higher throughput and accuracy. The rigid-body design paradigm aims at very stiff designs, which lead in an evolutionary way to increasingly heavier systems. Such systems require more and more power, such that this paradigm rapidly approaches the boundary of its scalability. An alternative paradigm is to design a lightweight machine with over-actuation and over-sensing, to deal with the resulting flexibilities. This paper presents a spatial feedforward method for over-actuated flexible motions systems, which aims at reducing the vibrations over the complete flexible structure during motion. The proposed method is experimentally validated on an industrial prototype and compared to mass feedforward and the standard input shaping technique.     

Adaptive control for enhancing tracking performances of flexible tendon–sheath mechanism in natural orifice transluminal endoscopic surgery (NOTES)

Tendon–sheath mechanism (TSM) has inherent advantages in the development of flexible robotic systems because of its simplicity, safety, flexibility, and ease of transmission. However, the control of TSM is challenging due to the presence of nonlinearities, namely friction, backlash-like hysteresis and the time-varying configuration of the TSM during its operations. Existing studies of TSM found in the literature only address tendon transmission under the assumption of fixed configuration and a complex inverse model of backlash is required. In order to flexibly use the system in a wider range of applications, the aforementioned nonlinear effects have to be characterized for the purpose of compensation. In this paper, we endeavor to address these issues by presenting a series of controller strategies, namely a feedforward control scheme under the assumption of known backlash-like hysteresis profile, and an adaptive control scheme to characterize the nonlinearities with unknown backlash hysteresis and uncertainties. The proposed control schemes do not require information of the tendon–sheath configurations, which is challenging to obtain in practice, in the compensation structures. In the absence of output position feedback, a simple direct inverse model-based feedforward has been used that efficiently reduce the tracking errors. The feedforward compensation does not require any complex algorithm for the inverse model. In the presence of output position feedback, a nonlinear adaptive controller has been developed to enhance the tracking performances of the TSM regardless of the random change in the tendon–sheath configurations during compensation. In addition, exact values of the model parameters are not required. They are estimated online during the operations. A dedicated experimental setup is introduced to validate the proposed control approaches. The results show that the proposed control schemes significantly improve the tracking performances for the TSM in the presence of uncertainties and time-varying configurations during the operations. There is a significant decrease of 0.0158 rad² (before compensation) to smaller value of 0.0012 rad² (use feedforward control) and 8.2815 × 10⁻⁵ rad² (use nonlinear controller) after compensation.     

Nonlinear identification and optimal feedforward friction compensation for a motion platform

In this study, we present a method of nonlinear identification and optimal feedforward friction compensation for an industrial single degree of freedom motion platform. The platform has precise reference tracking requirements while suffering from nonlinear dynamic effects, such as friction and backlash in the driveline. To eliminate nonlinear dynamic effects and achieve precise reference tracking, we first identified the nonlinear dynamics of the platform using Higher Order Sinusoidal Input Describing Function (HOSIDF) based system identification. Next, we present optimal feedforward compensation design to improve reference tracking performance. We modeled the friction using the Stribeck model and identified its parameters through a procedure including a special reference signal and the Nelder–Mead algorithm. Our results show that the RMS trajectory tracking error decreased from 0.0431 deg/s to 0.0117 deg/s when the proposed nonlinear identification and friction compensation method is utilized.     

Magnetic disturbance compensation for a reticle stage in a lithographic tool

The wafer stage and reticle stage in a lithographic tool, used to manufacture integrated circuits, operate at nanometer accuracy during a scanning motion. Magnetic interaction with the environment causes disturbance forces that result in a 5 nm moving-average filtered scanning error of the reticle stage.Stage motor magnet interaction with ferromagnetic material in the vicinity causes a position-dependent disturbance force, and eddy currents in electrically conductive metals cause a position-dependent damping force which is proportional with velocity.The total disturbance force can be estimated from control-loop signals, and by moving the stage in both directions both disturbance types can be distinguished from each other and recorded in compensation tables. Applying these tables as a feedforward reduces the scanning position error from 5 nm to 1.8 nm in a production reticle stage system.     

一种基于逐维扩展的前向网络训练方法

前向网络的快速训练问题是前向网络研究的一个非常重要的课题。本文针对一类n-维超立方体的分类问题（当为二分类问题时，这实际上是一个n-维Boole函数的神经网络实现问题），提出了一种基于逐维扩展的前向网络快速训练方法，将一个n个输入的大网络的各权训练问题转化为小网络逐维递归的扩展部分的参数训练问题，提高了网络训练的速度，实验结果表明了这种训练方法的有效性和可行性。     

A novel feedforward–feedback suboptimal control of linear time-delay systems

This paper presents a new approach for solving the optimal control problem of linear time-delay systems with a quadratic cost functional. In this approach, a method of successive substitution is employed to convert the original time-delay optimal control problem into a sequence of linear time-invariant ordinary differential equations (ODEs) without delay and advance terms. The obtained optimal control consists of a linear state feedback term and a forward term. The feedback term is determined by solving a matrix Riccati differential equation. The forward term is an infinite sum of adjoint vectors, which can be obtained by solving recursively the above-mentioned sequence of linear non-delay ODEs. A fast-converging iterative algorithm for this purpose is presented which provides a promising possible reduction of computational efforts. Numerical examples demonstrating the efficiency, simplicity and high accuracy of the suggested technique have been included. Simulation results reveal that just a few iterations of the proposed algorithm are required to find an accurate enough feedforward–feedback suboptimal control.     

Evaluation of fully fuzzy matrix equations by fuzzy neural network

In this paper, a new hybrid method based on fuzzy neural network for approximate solution of fully fuzzy matrix equations of the form AX=D$\mathit{AX}=D$, where A and D are two fuzzy number matrices and the unknown matrix X is a fuzzy number matrix, is presented. Then, we propose some definitions which are fuzzy zero number, fuzzy one number and fuzzy identity matrix. Based on these definitions, direct computation of fuzzy inverse matrix is done using fuzzy matrix equations and fuzzy neural network. It is noted that the uniqueness of the calculated fuzzy inverse matrix is not guaranteed. Here a neural network is considered as a part of a large field called neural computing or soft computing. Moreover, in order to find the approximate solution of fuzzy matrix equations that supposedly has a unique fuzzy solution, a simple algorithm from the cost function of the fuzzy neural network is proposed. To illustrate the easy application of the proposed method, numerical examples are given and the obtained results are discussed.     

Extended and Unscented Kalman filtering based feedforward neural networks for time series prediction

With the ability to deal with high non-linearity, artificial neural networks (ANNs) and support vector machines (SVMs) have been widely studied and successfully applied to time series prediction. However, good fitting results of ANNs and SVMs to nonlinear models do not guarantee an equally good prediction performance. One main reason is that their dynamics and properties are changing with time, and another key problem is the inherent noise of the fitting data. Nonlinear filtering methods have some advantages such as handling additive noises and following the movement of a system when the underlying model is evolving through time. The present paper investigates time series prediction algorithms by using a combination of nonlinear filtering approaches and the feedforward neural network (FNN). The nonlinear filtering model is established by using the FNN’s weights to present state equation and the FNN’s output to present the observation equation, and the input vector to the FNN is composed of the predicted signal with given length, then the extended Kalman filtering (EKF) and Unscented Kalman filtering (UKF) are used to online train the FNN. Time series prediction results are presented by the predicted observation value of nonlinear filtering approaches. To evaluate the proposed methods, the developed techniques are applied to the predictions of one simulated Mackey-Glass chaotic time series and one real monthly mean water levels time series. Generally, the prediction accuracy of the UKF-based FNN is better than the EKF-based FNN when the model is highly nonlinear. However, comparing from prediction accuracy and computational effort based on the prediction model proposed in our study, we draw the conclusion that the EKF-based FNN is superior to the UKF-based FNN for the theoretical Mackey-Glass time series prediction and the real monthly mean water levels time series prediction.     

基于线性曲率校正技术的低电压带隙基准源设计

采用电流求和取代传统电压求和的方法设计了一款低基准电压输出的带隙基准电压源电路,同时提出一种线性化P-N结正向导通电压（VBE）的温度曲率校正技术,保证了基准电压的低温漂和高精度。整个电路采用TSMC0.6μmBCD工艺设计实现,芯片面积为0.2mm2。在Cadence环境下使用Spectre对电路进行了模拟仿真,仿真结果表明：该基准电路可在低至1.1V的电源电压下正常工作;在-20℃～120℃温度范围内,温度系数为9.1×10-6/℃,PSRR为-78dB。在典型的1.5V电源电压下,基准输出电压可调节范围为0.165～1.25V。