A novel order-reduction technique for model predictive control(MPC) is proposed based on the singular value decomposition(SVD) of a pulse response circulant matrix(PRCM) of a concerned system. It is first investigated that the PRCM (in the limit) contains a complete information of the frequency response of a system and its SVD decomposes the information into the respective principal directions at each frequency. This enables us to isolate the significant modes of the system and to devise the proposed order-reduction technique. Though the primary purpose of the proposed technique is to diminish the required computation in MPC, the clear frequency decomposition of the SVD of the PRCM also enables us to improve the robustness through selective excitation of frequency modes. Performance of the proposed technique is illustrated through two numerical examples.

It is the purpose of this paper to make a preliminary study on the use of repetitive control to improve velocity accuracy by eliminating repetitive disturbances caused by machining inaccuracies of the axis of rotation location. If the control system can be intelligent enough to compensate for such machining errors, then one may be able to improve the accuracy of the velocity control, or alternatively, one may maintain the same accuracy and relax the machining tolerances required. This could decrease cost significantly. Experiments are performed testing repetitive control methods on a constant speed rolling operation testbed. The experimental results show very substantial decreases in the tracking error of the system. Spectral data of the output motion are given to demonstrate the attenuation of the disturbance frequencies and harmonics, related to the bandwidth being used. It is seen that the simplest form of repetitive control which is very easily implemented, can produce striking improvement in control system performance in such belt drive rolling operations, and the learning can be accomplished in a short time.

There are many demands for ship control system and many studies have been proposed. For example, if a ship diesel engine is operated by consolidated control with Controllable Pitch Propeller(CPP), the minimum fuel consumption is achieved, satisfying the demanded ship speed. For this, it is necessary that the ship is operated on the ideal operating line which satisfies the minimum fuel consumption. In this context of view, this paper presents a controller design method for a ship propulsion system with CPP by Linear Matrix Inequality(LMI) which satisfies the given $H_{\infty}$ control performance and robust stability in the presence of physical parameter perturbations. The validity and applicability of this approach are illustrated through a simulation in the all operating ranges.

The purpose of this study is to provide an integrated process control system for plasma etching. The control system is designed to employ neural network for the modeling of plasma etching process and to utilize genetic algorithm to search for the appropriate selection of control input variables, and to provide a control chart to maintain the process output within a desired range in the real plasma etching process. The target equipment is the one operating in DRAM production lines. The result shows that the integrated system developed is practical value in the improved performance of plasma etching process.

Motivated by a recent work, a fuzzy-power-controller(FPC) is designed for the relieving-horsepower control of output variable pump with electrical proportional valve and actually implemented on the industrial excavator. In order to calculate the output power of pump with input of FPC, a linear discrete time model of load system to pump is obtained and the result is applied to control the engine-pump coupled system by software without pressure and flow sensor. The FPC controls the engine and pump coupled system by relieving horsepower control according to the change of load and the running conditions in relieving horsepower control are selected by fuzzy inference engine. A case study is peformed through the construction of the control device and installation on the excavator. It shows that the relieving-horsepower control system with the FPC, as suggested in this paper, is superior to the conventional PID controllers. And also, the excavator, with the FPC, shows that the power-loss of the coupled system is reduced and the running speed of the hydraulic actuator is enhanced.

A neural network based control algorithm with fuzzy compensation is proposed for the automated adjustment in the production of electronic end-products. The process of adjustment is to tune the variable devices in order to examine the specified performances of the products ready prior to packing. Camcorder is considered as a target product. The required test and adjustment system is developed. The adjustment system consists of a NNC(neural network controller), a sub-NNC, and an auxiliary algorithm utilizing the fuzzy logic. The neural network is trained by means of errors between the outputs of the real system and the network, as well as on the errors between the changing rate of the outputs. Control algorithm is derived to speed up the learning dynamics and to avoid the local minima at higher energy level, and is able to converge to the global minimum at lower energy level. Many unexpected problems in the application of the real system are resolved by the auxiliary algorithms. As the adjustments of multiple items are related to each other, but the significant effect of performance by any specific item is not observed. The experimental result shows that the proposed method performs very effectively and are advantageous in simple architecture, extracting easily the training data without expertise, adapting to the unstable system that the input-output properties of each products are slightly different, with a wide application to other similar adjustment processes.

This paper presents an autopilot design method for STT missiles using the intelligent control technique and multiple controllers. The mixed $H_2/H_{\infty}$ control technique is applied for each controller design and the control gains are implemented by using the genetic searching algorithm. To facilitate automatic switching of multiple controllers under different operating conditions, an error based switching scheme is also combined with the multiple controllers at a higher level, which constitutes a hierarchical intelligent control system. It is shown via computer simulation that the proposed autopilot outperforms the conventional one.

In teleoperation, it is important for an operator to feel as if he really were in a distant place. To realize this objective, the various information from a remote site must be presented to the operator. Even though tactile information is very important to efficiently execute a task, it is not yet sufficiently provided for the operator. In this paper, we propose the new mechanism that can provide the more dexterous tactile information to the operator This device utilizing the electromagnetic force is designed to be compact and light enough to be attached to the fingerpad, and designed to be controlled continuously. The magnetic circuit is derived by the probable flux paths method in order to take forces at any given dimension. An optimization technique is also proposed to maximize the tactile force that humans can perceive under the same conditions. The objective function is formulated as maximizing displacements indented on the fingerpad, considering the mechanism of human tactile perception. The optimization formulation is subject to the geometric and rising temperature constraints in the coil. It is demonstrated that, by optimization, the tactile force increases by 24%, compared with that obtained from the initial design.

In this paper, the design and the implementation of a robust adaptive controller for trajectory tracking of robot manipulator is presented. The proposed control scheme ensures that tracking errors are converged to some boundaries in the presence of a state-dependent input disturbances as well as the ideal case without any prior knowledge of the robot manipulator parameters. The 3 DOF robot manipulator including actuator dynamics is used for the implementation of the proposed control scheme. The experimental results show that the proposed control scheme is valid for trajectory tracking of the robot manipulator.

Camera calibration is the process of determining the coordinate relationship between a camera image and its real world space. Accurate calibration of a camera is necessary for the applications that involve quantitative measurement of camera images. However, if the camera plane is parallel or near parallel to the calibration board on which 2 dimensional objects are defined(this is called "ill-conditioned"), existing solution procedures are not well applied. In this paper, we propose a neural network-based approach to camera calibration for 2D images formed by a mono-camera or a pair of cameras. Multi-layer perceptrons are developed to transform the coordinates of each image point to the world coordinates. The validity of the approach is tested with data points which cover the whole 2D space concerned. Experimental results for both mono-camera and stereo-camera cases indicate that the proposed approach is comparable to Tsai's method［8］. Especially for the stereo camera case, the approach works better than the Tsai's method as the angle between the camera optical axis and the Z-axis increases. Therefore, we believe the approach could be an alternative solution procedure for the ill -conditioned camera calibration.libration.

이 논문은 에나멜 전선의 제조공정의 자동화에 있어서 핵심역할을 하는 보빈의 자동인식을 위한 머신 비젼 시스템에 관한 것이다. 이 시스템의 역할은 컨베이어 라인의 팔레트 위에 놓인 보빈들의 영상을 CCD 카메라로 취득, 분석하여 보빈 형태, 색상, 제조공정번호 등의 다양한 정보를 추출하여, 전체 생산공정을 제어하는 주 컴퓨터로 보내는 일을 수행한다. 이 비젼 시스템은 개발된 후 에나멜 전선 생산공장에 설치되어 일정 시험기간을 거쳐 현재 성공적으로 운영되고 있다.

In developing an automated surface inspect algorithm, we have designed a hierarchical classifier using neural network. The defects which exist on the surface of cold mill strip have a scattering or singular distribution. We have considered three major problems, that is preprocessing, feature extraction and defect classification. In preprocessing, Top-hit transform, adaptive thresholding, thinning and noise rejection are used Especially, Top-hit transform using local minimax operation diminishes the effect of bad lighting. In feature extraction, geometric, moment, co-occurrence matrix, and histogram ratio features are calculated. The histogram ratio feature is taken from the gray-level image. For defect classification, we suggest a hierarchical structure of which nodes are multilayer neural network classifiers. The proposed algorithm reduced error rate by comparing to one-stage structure.

In this paper, an algorithm is proposed to detect the malfunction of plasma-etching characteristics using EPD signal trajectories. EPD signal trajectories offer many information on plasma-etching process state, so they must be considered as the most important data sets to predict the wafer states in plasma-etching process. A recent work has shown that EPD signal trajectories were successfully incorporated into process modeling through critical parameter extraction, but this method consumes much effort and time. So Principal component analysis(PCA) can be applied. PCA is the linear transformation algorithm which converts correlated high-dimensional data sets to uncorrelated low-dimensional data sets. Based on this reason neural network model can improve its performance and convergence speed when it uses the features which are extracted from raw EPD signals by PCA. Wafer-state variables, Critical Dimension(CD) and uniformity can be estimated by simulation using neural network model into which EPD signals are incorporated. After CD and uniformity values are predicted, proposed algorithm determines whether malfunction values are produced or not. If malfunction values arise, the etching process is stopped immediately. As a result, through simulation, we can keep the abnormal state of etching process from propagating into the next run. All the procedures of this algorithm can be performed on-line, i.e. wafer to wafer.

This paper discusses the development of a reactive power controller using digital signal processing. Digital Signal Processing is the technique of using digital devices to Process continuous signals or data, often in real-time. And DSP algorithms are associated with a discrete time interval between input samples. When one designs a digital filter, one can use a Laplace transform to determine the continuous time frequency response. The corresponding discrete time transform is called Z transform and depends upon discrete samples of the input spaced equally in time. The objectives of this paper are to minimize real power losses and improve the power factor of a given system. Also, the implementation of a direct-form non recursive filter on the TMS320C31 has been described. The application of this microprocessor-based controller using DSP on test system reveals its numerous advantages. Performance and features of the controller for the reactive power control are analyzed.

A conventional contact type brake system which uses a hydraulic system has mny Problems such as time delay response due to pressure build-up, brake pad wear due to contact movement, bulky size, and low braking performance in high speed region. As vehicle speed increases, a more powerful brake system is required to ensure vehicle safety and reliability. In this work, a contactless brake system of an eddy current type is proposed to overcome problems. Optimal torque control which minimizes a braking distance is investigated with a scaled-down model of an eddy current type brake. It is possible to realize optimal torque control when a maximum friction coefficient (or desired slip ratio) corresponding to road condition is maintained. Braking force analysis for a scaled-down model is done theoretically and experimentally compensated. To accomplish optimal torque control of an eddy current type brake system, a sliding mode control technique which is, one of the robust nonlinear control technique is developed. Robustness of the sliding mode controller is verified by investigating the braking performance when friction coefficient is varied. Simulation and experimental results will be presented to show that it has superior performance compared to the conventional method.

In this paper a simulation model for the powertrain control of gasoline engines with automatic transmission is presented. A modular programming approach has been pursued and the MATLAB/SIMULINK has been utilized as the programming environment. The engine/transmission system is analyzed in the object-oriented fashion whereby easy transferal of the modules, which represent physical parts or analysis subsystems, is guaranteed. Some mathematical models are adopted from the literature to compare the simulation results with the model and the experimental results in the literature. It is expected that the whole program or individual module constructed in this paper are useful for the automotive engineers in designing a new engine/transmission system and/or in modifying parts of existing systems.

From the power plant in a steel plant, environment pollutants such as $SO_x$, $NO_x$, CO and $CO_2$ are emitted by combustion reactions of the fuels which are by-product gases, oil and liquefied natural gas(LNG). To reduce the amounts of the pollutants, it is important to build a predictive model for the emission of the pollutants. In this paper, model that predict the amounts of generated pollutants for the used fuel is developed by using Gibbs free energy minimization method[1] with the temperature correction technique. For some data set, the calculation results from this model are compared with the real emission amounts of $SO_x$, $NO_x$, and the result of the calculation by both ASPEN PLUS which is a commercial simulation software. This model shows good results and can be applied to other power plants.

This paper presents a neural network model representing complex hydro-thermo-dynamic characteristics of a steam generator in nuclear power plants. The key modeling processes include training data gathering process, analysis of system dynamics and determining of the neural network structure, training process, and the final process for validation of the trained model. In this paper, we suggest a training data gathering method from an unstable steam generator so that the data sufficiently represent the dynamic characteristics of the plant over a wide operating range. In addition, we define the inputs and outputs of neural network model by analyzing the system dimension, relative degree, and inputs/outputs of the plant. Several types of neural networks are applied to the modeling and training process. The trained networks are verified by using a class of test data, and their performances are discussed.