All issues
Volume 41, Issue 1
Displaying 1-14 of 14 articles from this issue
- |<
- <
- 1
- >
- >|
-
Yongwoon CHOI, Kazuaki UENO, Megumi KAMIUCHI, Taketoshi IYOTA, Kazuhir ...2005 Volume 41 Issue 1 Pages 1-7
Published: January 31, 2005
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSA new image processing technique to apply an approximate plane of brightness has been proposed for inspecting defects of irregular pixels in a CCD image captured in proximate front of a LCD panel. The technique is capable of making automatic and speedy detection of defects, by reciprocally comparing the gradient of an approximate plane, which is derived from the brightness of each pixel in a small area divided into the identical section with a reference image. By exploiting the difference of the gradients in distinguishing the section associated with defects, there are a few characteristics as compared with the typical image processing techniques such as the differential and the template method. One is robust to noises that have effect upon the captured image from electric circuits constructing a camera itself. The other need not perform the template processing by a particular operator to emphasize a kind of defect in implementing image processing. Another is able to reduce the processing time because a captured image can be treated at a stretch without regard to kinds of defect. In this paper, we experimentally make it clear that the technique is able to detect not only a dot and a line defect of high variation, but also non-uniformity defects of comparatively low variation in brightness of pixels. Experiments with pseudo-defects represented in an actual LCD panel demonstrate that from the results obtained in 3colors×0.12seconds of processing time, the technique is able to be applied to a system for implementing defect-inspection of a certain uniformity level in real time.View full abstractDownload PDF (3510K) -
Masayuki TANAKA, Jun-ya TAKAYAMA, Shinji OHYAMA, Akira KOBAYASHI2005 Volume 41 Issue 1 Pages 8-15
Published: January 31, 2005
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSA cross-correlation envelope method which generates a model fitness from a reflected wave is used as pre-process for an image reconstruction. The range of the model fitness is normalized from zero to one because the effect of various attenuation is reduced by the corss-correlation envelope method. The image reconstruction method based on the micro-mirror model, or Phong model, for GPR is newly proposed. A subsurface image with tangential information of a boundary between a buried object and soil, which is useful to recoginize the shape of the object, is generated by the proposed method. measurement waves are classfied with antennas arrangement based on the position and the orientation of the micro-mirror. The model fitness generated from classfied waves are synthesised. Three types of synthesis method are introduced; an arithemetic average, a geometric average, and a harmonic average. In this work, experiments of steel plate detection are demonstrated to present a effectiveness of proposed method. As the result, inclination angles of buried steel plate are detected by proposed method with the tangential information.View full abstractDownload PDF (3573K) -
Satoshi ISONAGA, Satoshi HATTORI, Masayuki OKAMOTO, Shogo TANAKA2005 Volume 41 Issue 1 Pages 16-24
Published: January 31, 2005
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSDeveloping the monitoring system in a bathroom is important to prevent the accidents there. One of the method to prevent the accidents in a bathroom without violeting one's privacy is using an acoustic sensor. The authors proposed an idea of immersing a pipe, which an acoustic sensor is set inside, in the water. By this procedure, we can obtain not only a direct sound information from the people in the bathroom but also the water level information. That is, stationary wave depending on the water level will be formed inside the pipe and the frequency analysis of it enables the real time measurement of the level. But a thermometer is needed to the measurement of the water level in the proposed method.
To overcome this problem, the authors propose a new idea of using a pipe which has a narrow portion. That is, 2 types of the stationary wave are formed in the pipe. Therefore, we can obtain a direct sound information and the water level information without a thermometer.View full abstractDownload PDF (2096K) -
Takeshi MIZUNO, Minoru TAKEUCHI, Masaya TAKASAKI, Yuji ISHINO2005 Volume 41 Issue 1 Pages 25-30
Published: January 31, 2005
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSA mass measurement system using the self-excited oscillation of a relay feedback system was proposed. The system contains an on-off relay with dead zone and switches force acting on the object in relation to the position. An analytical study showed that the mass of the object is determined from the time interval measurement of the on-state and off-state periods. An apparatus was developed for experimental study. It uses a voice coil motor for generating force, and a pair of photo interrupters for detecting the switching positions. The effects of system parameters on measurement accuracy were studied experimentally. Under the tuned conditions, the measurement errors were within 0.2%. The effects of damping on measurement accuracy were also studied experimentally.View full abstractDownload PDF (2582K) -
Toward Nano Order ControlAtsuhiko OKAO, Masao IKEDA, Ryoichi TAKAHASHI2005 Volume 41 Issue 1 Pages 31-36
Published: January 31, 2005
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn recent developments of nano technology, there are demands of controlling mechanical systems in nano meter orders. For this purpose, we need a plant model which describes the behaviors in the nano orders. In system identification of continuous-time plants, we commonly use digital data converted from analog signals. Since digital variables are of finite wordlengths, the digital data contain quantization errors. Such errors may deteriorate identification accuracy significantly, especially in the case of mechanical systems where both large movement and precise positioning are required. To overcome this problem, the present paper proposes estimation of the quantization errors and thus true sampled analog signals to improve identification accuracy. Estimates of quantization errors and system parameters are computed simultaneously. By simulation, it is illustrated that accurate results can be obtained by the proposed method.View full abstractDownload PDF (1582K) -
Masahiro OYA, Makoto WADA, Toshihiro KOBAYASI2005 Volume 41 Issue 1 Pages 37-45
Published: January 31, 2005
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSBy using a state observer a new robust trajectory tracking control scheme is developed in this paper for flexible joint manipulators. A precompensator including an integrator is introduced in order to ensure asymptotical stability of tracking error between link angles and constant desired trajectories. Then, to ease the development of a controller achieving good robust tracking performance, a new representation of flexible joint robot dynamics with the precompensator is introduced. The developed controller based on the new representation has the following features; i) The configuration of the proposed controller becomes very simple compared with regressor based adaptive controllers. ii) The Euclid norm of tracking errors arrives to any small closed region with any convergent rate by setting design parameters. Especially for the desired trajectories converging to constant ultimate values, tracking errors converge to zero.View full abstractDownload PDF (1950K) -
Sadaaki KUNIMATSU, Takao FUJII2005 Volume 41 Issue 1 Pages 46-52
Published: January 31, 2005
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe robust servo system with robustness of output trajectories is analyzed and synthesized in the design of model reference ILQ (MR-ILQ) servo systems with the potential to achieve tracking robustness. Our purpose is to design a robust servo system with tracking robustness in that its output trajectory remains in a neighborhood of the nominal trajectory when the dynamics of the system is perturbed. For this purpose we first propose a robust L∞ performance measure for a fixed input in order to evaluate the worst-case value of the maximal amplitude of the error output response. By analyzing some related convex optimization problem described by linear matrix inequalities (LMIs), we next present a specific method to calculate this measure for the MR-ILQ servo systems. Based on this analysis, we then formulate a synthesis problem of minimizing this measure, and derive an LMI-based iterative algorithm to solve this synthesis problem. Finally we illustrate a numerical example to show the effectiveness of our results.View full abstractDownload PDF (1342K) -
Application to Design of Flight Vehicle ControlIzumi MASUBUCHI, Jun KATO, Masami SAEKI, Atsumi OHARA2005 Volume 41 Issue 1 Pages 53-60
Published: January 31, 2005
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper is concerned with a synthesis method of gain-scheduling controllers with utilizing descriptor representations and its application to design of flight vehicle control. By representing a linear parameter-varying plant in a descriptor form, parameter-dependent LMIs to seek a gain-scheduling controller are formulated so that the LMIs have simple structure with respect to the parameter. A gain-scheduling controller is computed directly via an explicit formula in terms of the variable of LMIs for descriptor LPV systems. This synthesis procedure is applied to design of a flight vehicle control with large variation of dynamics and multiple control specifications. It is shown that the obtained gain-scheduling controller attains all the specifications and control performance is improved by employing parameter-dependent Lyapunov matrices.View full abstractDownload PDF (1725K) -
Takao NAMBU2005 Volume 41 Issue 1 Pages 61-66
Published: January 31, 2005
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSWe study stabilization of a class of linear parabolic boundary control systems. The control system in this paper admits a Riesz basis corresponding to a pair of coefficient linear differential operators (L, τ). The problem has been extensively studied by several authors. Most important is how we cope with the boundary control terms. We propose a new algebraic method of stabilization, which is regarded as the third algebraic one by the same author. An advantage is that this method is a little simpler and more readable than the exixsting ones, since the argument is based on a system of differential equations with no boundary input. Thus the argument owns some property in common with the one in finite-dimensional control systems.View full abstractDownload PDF (938K) -
Takafumi SASAKAWA, Jinglu HU, Kotaro HIRASAWA2005 Volume 41 Issue 1 Pages 67-74
Published: January 31, 2005
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn an ordinary artificial neural network, individual neurons do not have any special relations with input patterns. That is, an ordinary neural network has only learning capability, but does not have the capability of function localization. However, according to Hebb's cell assembly theory about how the brain worked, it is suggested that it has function localization in the brain, which means that specific groups of neurons are activated corresponding to certain sorts of sensory information the brain receives. On the other hand, it is also reported that the cerebellum and cerebral cortex in the brain are specialized in supervised and unsupervised learning paradigms, respectively. Inspired by both Hebb's cell assembly theory, and the basic learning paradigms in the brain, this paper presents a self-organizing function localization neural network (FLNN). The proposed self-organizing FLNN consists of two parts: the main part and the control part. The main part, corresponding to the cerebellum of brain, is an ordinary 3-layered feedforward neural network, but each hidden neuron contains a signal from the control part, controlling its firing strength. The control part, corresponding to the cerebral cortex of brain, consists of a self-organizing map (SOM) network whose outputs are associated with the hidden neurons of the main part. Trained with an unsupervised learning algorithm, the SOM control part extracts structural features of input space and controls the firing strength of hidden neurons in the main part. And the main part realizes an input-output mapping by using supervised learning. In this way, the self-organizing FLNN realizes the capabilities of both function localization and learning. Numerical simulations show that the self organizing FLNN has superior performance to an ordinary neural network.View full abstractDownload PDF (2032K) -
A Parameterization of the Simultaneous Stabilizers for Simultaneously Stabilizable PairsOsamu KANEKO, Kazuya MORI, Kyoko YOSHIDA, Takao FUJII2005 Volume 41 Issue 1 Pages 75-84
Published: January 31, 2005
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn this paper, we address synthesis of reliable control systems based on simultaneous stabilization problem in a behavioral framework. Firstly, we provide a necessary and sufficient condition for a control law to stabilize given two plants simultaneously. Nextly, by using this condition, we give a parameterization of all of the simultaneous stabilizing controllers for given two plants. By focusing on the single-input single-output case, we also give a condition that enables us to check whether given two plants are simultanously stabilizable in terms of poles and zeros of the transfer functions, which is a generalization of the results obtained in the standard system theory. Moreover, we give an illustrative example in order to show the validity of our results.View full abstractDownload PDF (2672K) -
Tadashi EGAMI, Koichiro MOTIZUKI, Satoshi NAKAZAKI2005 Volume 41 Issue 1 Pages 85-87
Published: January 31, 2005
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSWe propose a new path control for vehicles using the time-state control form with the rotational and expansionary coordinate transformations. We can realize a desirable path control to arbitrary desired curve paths by the proposed method.View full abstractDownload PDF (489K) -
Kiminao KOGISO, Kenji HIRATA2005 Volume 41 Issue 1 Pages 88-90
Published: January 31, 2005
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper shows how to construct a reference governor whose implemented data size can be adjusted with an integer parameter ν, ensuring fulfillment of a pointwise-in-time constraint. Using a technique “blocking”, a management method is established that changes the external reference every νTs period, where Ts denotes a sampling period of a constrained system. The constraint fulfillment for an infinite time is achieved by satisfying a terminal condition described by a maximal output admissible set. The reference governor is finally obtained with an explict solution to the convex quadratic programming problem. A numerical evaluation demonstrates its effecitveness for constraint fulfillment and data compression.View full abstractDownload PDF (610K) -
Koji IKENOYA, Ryota YAMASHINA, Masayuki HARA, Tetsuro YABUTA2005 Volume 41 Issue 1 Pages 91-93
Published: January 31, 2005
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIterative learning control can make the response of the system improve with the way of repetition trials of the control. However, when this control method is applied to a non-minimum phase system, there is the possibility of no improvement by the learning control. In this paper, we modify this learning control method and it is applied to vibration control of flexible arm system which is the non-minimum phase. Results show good improvement with the modified control algorithm.View full abstractDownload PDF (500K)
- |<
- <
- 1
- >
- >|