All issues
Volume 32, Issue 6
Displaying 1-25 of 25 articles from this issue
- |<
- <
- 1
- >
- >|
-
Takeshi HATSUZAWA1996 Volume 32 Issue 6 Pages 791-794
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSPrecise measurements of focused electron beam diameters are usually performed by measuring the beam current distribution using a mechanical knife edge line scan. In the measurements, the positioning resolution of the knife edge affects the beam diameter accuracy, which is not so highly reliable because of the difficulty of the installation of a precise measurement instrument in the vacuum chamber of scanning electron microscopes (SEMs). Against this backdrop, a compact laser interferometer with a piezodriven micro mechanical scanner is installed in an SEM to realize accurate knife edge positioning. The interferometer has a resolution of 0.4μm (1/1600 of He-Ne laser wavelength - 0.633μm), which corresponds to the sampling interval of the current distribution measurement. From experimental results for different acceleration voltages and probe currents, a maximum standard deviation of the measurement of 3nm was obtained for beam diameters of a couple nm through dozens of nm.View full abstractDownload PDF (1941K) -
Tohru KATO, Makoto TAKAHASHI, Kazutaka MITOBE, Masahiro KIMURA, Yuichi ...1996 Volume 32 Issue 6 Pages 795-801
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSVirtual reality technology provides a new role of human-machine interface. In the virtual space, it is important that we cannot only perceive a spatial environment but also manipulate objects as we will do in the real space. In order to realize a virtual environment, it is necessary to investigate the relation between sensory system and motor system. Because it is natural that we usually interact with anything in the real space as well as we perceive the real space passively.
In this paper, we investigate the influence of visual stimulus on the hand movement induced visual stimulus. We use two visual stimulus in the experiments. One is a “Target” which subject is instructed to hold. Another is a “Reference point” that is presented as an additional stimuli with a “Target”. These two visual stimulus cannot be touched because of mirror images.
We assessed the pointing error to the target as the accuracy of space perception. The results revealed that in the case of presenting a reference point, the pointing error becomes smaller than in the case of presenting only target. Especially, this tendency is more effective on the target near the reference point.
Our findings suggest that visual stimulus near the target give the information to perceive more accurately the target position.View full abstractDownload PDF (2222K) -
Koji NAGANO, Hiroaki NIITSUMA1996 Volume 32 Issue 6 Pages 802-810
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSVelocity of guided waves which propagate along a subsurface fracture is estimated by means of a wave detection technique based on time-frequency representations. Propagation characteristics of the guided waves depend on not only geometric features of the interface but also mechanical properties on it. These interface conditions sensitively affect the propagation characteristics because their energy concentrates on the interface and they propagate along the interface. Measurement of the guided waves, therefore, has an advantage to characterize the interface condition. This energy distribution, however, makes it difficult to use an array sensor in velocity measurement. Number of sensors which can be arranged near the interface is limited, in particular, in a field measurement.
We examine a wave detection technique based on the time-frequency representations. The time-frequency representations combine time-domain and frequency-domain analyses to distinguish the guided waves from other waves and detect an arrival of each frequency component. This technique is applied to the estimation of the guided waves trapped in an artificial subsurface fracture. Specific patterns of both a wave source and detected guided waves are analyzed in the time-frequency representations. A result by the time-frequency representation analysis shows a typical dispersion pattern of the fundamental symmetric mode of the guided waves.View full abstractDownload PDF (2580K) -
Kiyomi HORIKOSHI, Yoshiyuki SHIRAKAWA1996 Volume 32 Issue 6 Pages 811-819
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper describes an X-ray radioscopy system which has been developed for on-line welding inspection of large diameter steel tubes incorporating X-ray image integration technique as well as simultaneous control method for X-ray tube voltage and current. The simultaneous control of X-ray voltage and current features regulation of X-ray intensity which has permeated through steel tubes of varying wall thickness. This system censures stable statistical error range which leads to clear image. Also an X-ray image integrating technique which places emphasis on the contrast of defects against background has been developed. These technologies feature stable integration process under tubes' fluctuating traveling speed in a mill. By utilizing the proposed methods, those images involving defects are integrated 20 to 30 times to reduce statistical errors and processed into clear and high contrast images. Our on-line and off-line tests show penetrameter sensitivity within 1.0 to 1.4% at the welding thickness ranging from 10 to 40mm. This X-ray radioscopy has been in operation at a large diameter steel tube mill and has contributed to improved quality control of the tube products.View full abstractDownload PDF (2007K) -
Evaluation of Its Reliability on The Fixed Point of Palladium by Using The Crucible and Wire MethodsTokio HAMADA, Haruki YAMASAKI1996 Volume 32 Issue 6 Pages 820-825
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe reliability on the fixed point of Palladium was evaluated by examination using the crucible and wire methods. As the results, the temperature of the fixed point was proved to be independant of realizable technique with the methods, but remarkably depends on an internal atmosphere, especially the effect of Oxygen in air, surrounding two samples (ingot or wire) for the fixed point. It is found that the temperature of the fixed point in argon is 1.7°C higher than that of the fixed point in air.View full abstractDownload PDF (1706K) -
Yasuhide KOBAYASHI, Toru ASAI, Shinji HARA1996 Volume 32 Issue 6 Pages 826-834
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn this paper, we propose a method to establish the robust performance based on a parameter dependent controller under the parameter estimation errors.
We first propose a model of the parameter which represents the estimation error less conservatively. A sufficient condition for the existence of a controller is derived based on the scaled maximum singular value condition, and it is given by LMIs (Linear Matrix Inequalities) with a BMI (Bilinear Matrix Inequality) constraint. The relation of estimation accuracy and performance level is also investigated by a numerical example. Furthermore, we introduce a bilinear transformation to represent the situation that the accuracy of the estimation depends on the value of the parameter, and it is shown that the controller can be constructed by a simple modification for the plant.View full abstractDownload PDF (1651K) -
Robust Control of Plants with Variable Operating ConditionsEiichi MURAMATSU, Masao IKEDA, Kunihiro MATSUMOTO1996 Volume 32 Issue 6 Pages 835-843
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSStabilization of a multi-input multi-output plant with variable operating conditions is considered. The plant is described as a linear interpolation of proper stable coprime factorizations of transfer matrices of two representative models which are defined at two representative operating points. This paper considers a certain set of interpolated models in which we expect the real plant is included. Stabilization of all interpolated models in the set by an interpolated controller or a fixed controller is considered. The problems are reduced to certain H∞ control problems. Sufficient conditions for stabilizability are obtained and expressed in terms of LMIs (linear matrix inequalities) in the state space.View full abstractDownload PDF (1242K) -
Masaru KOGA, Kotaro HIRASAWA, Masanao OBAYASHI, Jun-ichi MURATA1996 Volume 32 Issue 6 Pages 844-853
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSA method is proposed for controlling of chaotic phenomena on Universal Learning Network (U.L.N.).
U.L.N. is a network which consists of arbitrary nonlinear functional nodes and arbitrary delay elements between the nodes, and can model and control various large-scale complicated systems like industrial plants as well as economical, social and life phenomena.
The chaos control method proposed here is a novel one. Generation and die-out of chaotic phenomena are controlled by changing Lyapunov Number of U.L.N., which is accomplished by adjusting U.L.N. parameters so as to minimize a criterion function that is the difference between the desired Lyapunov Number and its actual value.
A gradient method is developed for optimizing the parameters. The method utilizes second order derivetives and a method for calculating the second order derivetives is also proposed which is applicable to a wide range of systems with recurrent dynamics.
Nevertheless, it is difficult to optimeze the parameters sucusessfully by using the gradient method only because the Lyapunov Number is very sensitive to changes in the parameters and the criterion function is not convex. Therefore, a random search is combined with the gradient method.
Controlling of generation and die-out of chaotic phenomena are realized in simulations. In the simulations, sigmoid functions are adopted as U.L.N. node functions. Therefore, the U.L.N. is a fully recurrent neural network. And the weights, the thresholds and the slopes of the sigmoid functions are optimized.
The chaos control method could be the first step to a new paradigm of control system which makes use of chaotic phenomena positively.View full abstractDownload PDF (1566K) -
Toshiharu SUGIE, Kenji FUJIMOTO1996 Volume 32 Issue 6 Pages 854-861
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper is concerned with the use of freedom in the exact linearization technique. First, the freedom of the change of coordinates and nonlinear state feedback in the exact linearization is clarified. Second, the relation between the original nonlinear system and the corresponding linearized system is examined to see the effect of the use of freedom in the change of coordinates. Finally, we give numerical examples with simulation to show how to use the freedom.View full abstractDownload PDF (1392K) -
Kou YAMADA, Keiji WATANABE1996 Volume 32 Issue 6 Pages 862-870
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn this paper, we consider a design method of the filtered inverse system for time-invariant system, which produces the input signal to the system approximately. The inverse system must be stable in practical application. It is well known that the inverse system is unstable if the system has invariant zeros in right half plane. In conventional design method of filtered inverse system for the system with some right half plane invariant zeros, we factorize the system into the product of the inner and outer function, and constract the filtered inverse for the outer function of the system. If the system is strictly proper, we must solve the generalized Riccati equation based on the descriptor form. This requires complex numerical computation.
The purpose of this paper is to present a new state-space design method of stable filtered inverse system for the strictly proper system with some right half plane invariant zeros without solving the generalized Riccati equation. We constract the preliminary filtered-inverse system by usual method and derive a pseudo proper plant from it. Then the problem is reduced to the problem for proper system. This method makes the computation very simple.View full abstractDownload PDF (1087K) -
Seigo SASAKI, Kenko UCHIDA1996 Volume 32 Issue 6 Pages 871-879
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn this paper, we consider H∞ output feedback control problems for bilinear systems. First we derive two types of H∞ output feedback controller via differential game approach. The controllers are characterized in terms of the solutions to two Riccati inequalities depending on the state of generalized plant or controller. We propose an algorithm to solve the. Riccati inequalities via specifying a domain of the state and solving constant coefficient Riccati inequalities. The proposed algorithm includes an evaluation method for the domain of internal stability. Based on this algorithm, finaly we present a design example of artificial rubber muscle actuator control system.View full abstractDownload PDF (1777K) -
Akihiko MATSUSHITA, Takeshi TSUCHIYA1996 Volume 32 Issue 6 Pages 880-886
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper describes on-line preview planning for a desired signal of a 3-wheeled vehicle.
We have proposed to combine on-line planning for a desired signal and control system design. In this method, the planner is designed as a part of the control system by traditional design methods for control system design. In off-line planning, the control system reduces the influence of disturbance, parameter perturbation and initial value. In on-line planning, both the planning for the desired signal and the control system can reduce the influence of disturbance etc. That is, on-line method can reduce the influence better than off-line planning by changing the desired signal with the state of the control system.
Control system performance can be improved by utilizing the future desired signal and/or the disturbance signal in advance. Such a control is called “PREVIEW CONTROL”. Preview control can't be utilized for on-line planning because the future desired signal can be varied with the future state of the control system. However, in the proposed “PREVIEW PLANNING” the future information of the restriction of the desired signal is utilized instead. The preview planning can prevent the controlled variables from spilling over the permitted domain better than without preview planning. In on-line planning a feedback loop is constructed and in preview planning feedforward is added to it. In the motion of 3-wheeled vehicles, the domain within which the vehicle can move is usually restricted by various obstacles etc. Where the information of the obstacles within a certain distance can be given by the sensor, on-line planner can utilize the future information of the restriction in advance. The vehicle can start to avoid the obstacles earlier than without preview planning. We show the simulation results in order to clarify its effectiveness.View full abstractDownload PDF (1048K) -
Ikuro MIZUMOTO, Mingcong DENG, Zenta IWAI1996 Volume 32 Issue 6 Pages 887-895
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper deals with a systematic and quantitative design method of a parallel feedforward compensator (PFC) for MIMO plants which virtually realizes an almost strictly positive real (ASPR) augmented plant. At present, development of design methods of such PFCs is requested from various control techniques based on output feedback control schemes including the simple adaptive control and so on. The design procedure of the proposed PFC is quite simple and easy for installation because the ASPR-ness of the augmented plant is confirmed by adjusting the coefficients of Hurwitz polynomials and one small positive parameter δ in the PFC. The effectiveness of the proposed method is confirmed by applying it to a 2-input/output unstable plant.View full abstractDownload PDF (1297K) -
Wennong ZHANG, Akihiko MATSUSHITA1996 Volume 32 Issue 6 Pages 896-903
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSUntil now, H∞ control theory has been aided by small gain or norm-bounding techniques to supply robust stability conditions. Many practical applications of robust feedback control involve constant real parameter uncertainty, whereas small gain or norm-bounding techniques guarantee robust stability against complex, frequency-dependent uncertainty, thus entailing undue conservatism and unnecessarily sacrificing performance. In thiss paper, we consider the robust stability by the Hurwitz criterion which is sufficient and necessary, and robust performance by satisfying an H∞ sensitivity constraint for an SISO minimum phase plant having parametric perturbation. We design the compensator so that the complementary sensitivity function is a low-pass filter function for an assumed nominal plant which does not necessarily exist in the family of real models. Thus the compensator robustly stabilizes the closed-loop system for a plant with large parametric variation. Furthermore, we obtain satisfactory performance in the low-frequency domain. Robust stability criteria, robust performance criteria and a number of assumptions are imposed. The satisfaction of these assumptions guarantees the existence of a proper compensator which satisfies mixed criteria.View full abstractDownload PDF (1377K) -
When Decelerating on a Downhill GradeHiroshi TAKAHASHI, Kouichi KURODA, Masayuki YASUOKA1996 Volume 32 Issue 6 Pages 904-911
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper presents a new methodology for controlling vehicular powertrain systems such as an automatic transmission. It usually takes about 500 milliseconds to complete gear shift control after a driver shifts gears. The control system proposed here is designed to infer the driver's intention before the person executes a gear shift. Inputs to the system therefore include both shifting operations and the driver's intentions. The identification model for inferring the driver's intentions is the focus of this paper. Quantitative and qualitative input variables are estimated by interpretative structural modeling (ISM), and the model is divided into a quantitative submodel and a qualitative submodel. Special attention is given here to the former submodel. The production rules used to infer the driver's intention to decelerate are generated experimentally by the iterative dichotomizer 3 (ID3) method. A comparison is made with a conventional control system to show the performance improvements obtained with the proposed system. The results verify the effectiveness of factoring the inference of the driver's intentions into powertrain control systems.View full abstractDownload PDF (3220K) -
Keiko SHIMIZU, Masashi NAKAMOTO, Takashi KOKUBO, Hitoshi TANABE1996 Volume 32 Issue 6 Pages 912-920
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSGeneralized Predictive Control (GPC) has been applied to develop an advanced control algorithm for a long dead time process in a thermal power plant. A precompensator is utilized to cope with process nonlinearity. Feedforward control is also achieved in the precompensator to reduce the effect of measurable disturbance change. Dual-rate sampling structure is introduced so that a feedforward control is carried out faster sampling rate than a feedback compensation achieved by GPC. The controller is experienced for a NOx decomposition process of a combined cycle power plant on commercial operations. Test results show that GPC satisfies requirements of actual process control and provides successful performance.View full abstractDownload PDF (1554K) -
Leo NAGAMATSU, Koichiro DEGUCHI1996 Volume 32 Issue 6 Pages 921-929
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper proposes a performance estimation method for multithreaded processor systems. Multithreaded processors are known as effective memory latency tolerating scheme with multiple hardware contexts in a processor. It is important to estimate the processor working ratio for comparing various configurations. The proposed method is based on, firstly, assuming independent cache miss occurrences, and modeling the processor status with number of runnable threads and status of the thread in the execution stage, then employing a probability transition matrix between all possible processor statuses. Our method drastically reduced the amount of computation required to estimate the working ratio than traditional simulations.View full abstractDownload PDF (1285K) -
Iko MIYAZAWA, Takashi SEKIGUCHI1996 Volume 32 Issue 6 Pages 930-937
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn this paper, the application of Petri nets is restricted to Sequential Control Systems, which are event-driven type and discrete systems possessing asynchronous and parallel properties. We have already proposed a method to model such Sequential Control Systems by Petri nets with place invariants, and also a method to divide a large Petri net to some small sub nets using the same place invariants. According to the structural character of Sequential Control Systems, the Petri net models have to have place invariants and transition invariants. In this paper, we discuss place invariants of Petri nets and the related problems. First, we prove that a net modeled by Petri nets with place invariants can be divided using the method of partition by the same place invariants. Then, the conditions of the properties holding between the whole and sub nets about covering by place invariants and liveness are clarified. Finally, a method to decide the reachability of a whole net through its sub nets will be introduced. Be based on these results, it can be concluded that the analysis of Petri net models of Sequential Control Systems is able to be treated through the analysis of their sub nets.View full abstractDownload PDF (1234K) -
Satoshi ITO, Yasuko NAKAYAMA, Hiroyuki MIZUTANI1996 Volume 32 Issue 6 Pages 938-947
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper proposes a model-based approach to generating graphical explanations of high-level specifications for plant control. The specifications are explained by a symbolic simulator which generates a plant animation. The animation corresponds directly to images of machines' action which the designers have in their minds, so that the designers can easily confirm the accuracy of their specifications.
Many kinds of knowledge about a plant are needed to generate the animation. This knowledge includes machine structures, machine actions, functions, materials, and so on.
The knowledge about functions is the most important for the plant animation, as it allows the symbolic simulator to reason about operations on the materials. In particular, when a plant deals with solid materials, it is difficult to represent this knowledge, because the machines used in such a plant usually have some functions which depend on machines' actions and plant conditions. To solve this problem, we have developed a framework to represent such functions and their relations instead of using a global behavior model.View full abstractDownload PDF (5693K) -
Nobuo SANNOMIYA, Koji NAGANO, Masami KURAMITSU1996 Volume 32 Issue 6 Pages 948-956
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSA mathematical model for schooling behavior of three fish in the two-dimensional unbounded space is considered. The model is represented by a system of nonlinear differential equations. The model parameters are estimated from water tank experiment data. It is shown by simulation and system analysis that some qualitatively different patterns of their behavior are obtained, depending on the values of the system parameters.
A simulation result shows that three fish swim along closed circular trajectories or unclosed ones (called flower pattern). An analytical investigation is also done on the stability of the steady states of their behavior. Consequently, the circular motion is a stable steady state for most cases of the model parameters. The motion drawing flower pattern is stable only for the case where the circular motion becomes unstable. Moreover, by changing the parameters, a transition from a circular motion to a flower pattern motion is realized.View full abstractDownload PDF (1274K) -
Haruhiko TAKASE, Kazutoshi GOHARA, Yoshiki UCHIKAWA1996 Volume 32 Issue 6 Pages 957-966
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSWe have already demonstrated that, by regarding the time variable external input of finite length as one cycle of the wave form of a periodic function, RNN's can be described as forced vibration systems. It was also pointed out that to explain behaviors of RNNs we must consider attractors created in the state space of RNNs corresponding to training patterns. This paper describes our recent systematic simulations to investigate into behaviors of additive nets, typical models of RNNs, In the case where multiple combinations of simple wave forms of limited length are employed as time sequential training patterns. The results showed that (1) a cyclic trajectory is created in the state space of the RNN corresponding to each of training patterns, and (2) transition time from one cyclic trajectory to another induced by switching from one input pattern to another can be shortened remarkably by selecting appropriate combinations of training patterns.View full abstractDownload PDF (2100K) -
Yoichi NAGAO, Hideaki OHTA, Hironobu URABE, Shin-ichi NAKANO, Sadatosh ...1996 Volume 32 Issue 6 Pages 967-974
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper discusses an autonomous distributed control system to overcome the deficiencies in robustness, flexibility, and set-up speed of traditional centralized control systems. In order to implement the autonomous distributed control necessary for intelligent manufacturing systems (IMS), we propose a model and a method to realize this model. The model consists of numerous autonomous agents, and a field where these agents can exchange information for cooperation. The autonomous distributed control system model was applied to an automated guided vehicle (AGV) system. Here, each agent was modelled with Petri net and the communication among agents was presented using message passing. Also, numerical simulations were conducted to evaluate the effectiveness of our model.View full abstractDownload PDF (2252K) -
Hiroshi ITO, Hiromitsu OHMORI, Akira SANO1996 Volume 32 Issue 6 Pages 975-977
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper shows that constantly-scaled nonlinear H∞ control is a solution to robust stabilization and robust performance problems of uncertain nonlinear systems under structured perturbation. The relationship between robust asymptotically stabilization and robust H∞ performance problems is also clarified.View full abstractDownload PDF (347K) -
Kenji SUGIMOTO, Yi LIU1996 Volume 32 Issue 6 Pages 978-980
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper proposes an alternative scheme for discrete-time adaptive control systems. Instead of the conventional adjusting law of recursive least square (RLS) type, the proposed scheme makes use of a recursive QR factorization technique. Stability of the obtained adaptive system can be shown by clarifying its relationship to the RLS adjusting law.View full abstractDownload PDF (363K) -
Toshinori FUJITA, Weimin PAN, Tatsuya MATSUNAMI, Toshiharu KAGAWA1996 Volume 32 Issue 6 Pages 981-983
Published: June 30, 1996
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper has clarified the dynamic disturbance characteristics of a control valve positioner using our proposed model. Obtained results lead to the conclusion that time constant of the diaphragm room pressure influences disturbance characteristics the most.View full abstractDownload PDF (415K)
- |<
- <
- 1
- >
- >|