-
Deddy KURNIADI, Kin-ichi KOMIYA
1996 Volume 32 Issue 3 Pages
281-289
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
This paper describes a method for estimating the internal distribution of electrical impedance of a compound object from the potential data on the boundary. The boundary potential are measured while a set of currents are applied to a number of electrodes on the periphery of the object. The current data are also presented to a finite element (FE) model of the object, with an initial guess of internal impedance distribution, and then the potential distribution is calculated. An objective function is introduced to define the error between the boundary potential of the FE model and that of the real object. The impedance parameters of the FE model are iteratively modified so as to minimize the objective function, using the Newton-Raphson method. To validate the approach, a computer simulation and an experiment were performed in which resistive-capacitive materials were considered.
View full abstract
-
Masayoshi TSUCHIYA, Takeshi KASAI, Masahiko HAYASHI, Yasunobu IWASAKA, ...
1996 Volume 32 Issue 3 Pages
290-296
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
Development has been done to fabricate a suitable aerosol sonde for balloon observation. This study aims to optimize sensitivity, weight, and dimension of the sonde. A 5.5kg, 25×25×15cm
3 sonde, which can detect as small as 0.23μm (PSL Sphere) in diameter, was realized. At low pressure (5∼100hPa) in stratosphere, a sampling flow rate of aerosol always decreases due to fluid mechanical problem of pumping. To overcome this difficulty, a gear pump, of which revolution speed was controlled by the environmental pressure, was adopted and it enabled the flow rate to keep at about ±5%, from 1013hPa to 5hPa in pressure. In stratosphere, the aerosol is generally composed of sulfuric acid mist. For the measurement of such a mist, the software, suitable to a personal computer, on basis of refractive index was developed, in order to correct the response curve of various kind of mist. This correction verified that reasonable experimental results were obtainable through the several field tests.
View full abstract
-
Shunji HIROKAWA
1996 Volume 32 Issue 3 Pages
297-305
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
While many studies on the anterior cruciate ligaments' (ACL) tensile characteristics have been done, the non-uniform distribution of strain on the fibers within the ACL has not been fully studied. Highly variable deformations of the ligaments (which are not uniaxial structures such as a bundle of separate fibers) could not be adequately quantified by one-dimensional and/or localized measurements.
This study employed an alternative approach: The ACL was treated as a biaxial structure, thereby introducing the change in shape and strain distribution over the entire surface. The stress analysis were carried out using the finite element method for incompressive hyper-elastic material to verify the previously performed experimental results for a pseudo ligament (rubber). The non-linear stiffness relation between the node force and the node displacement for a special triangular element, was introduced. The Newton-Raphson method was used to solve the nonlinear equations associated with finite element representations of the pseudo ligament.
Both the previous measurement and the present theoretical results were in good agreement. Special attention was paid for the distribution of longitudinal strain as a function of the knee flexion angle. Strain distribution varied, even along the fiber direction, and large strain gradients were observed in the regions near the bone attachments. Furthermore, the following values were obtained as a function of knee flexion: Resultant pull force on the tibial insertion, and change of strain distribution pattern by anterior-posterior displacement of the tibia.
View full abstract
-
Eiichi MIYAZAKI, Kensho OKAMOTO
1996 Volume 32 Issue 3 Pages
306-312
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
Recently a high speed pulsed light generator system has been developed for the optical analysis or imaging of fast processes in mechanical oscillations, turbulent velocity fluctuations in gaseous or liquid flows and plays an important role in visualizing the time dependency of rapidly changing states and processes. In this paper, a pulsed light system for stroboscopic photography that makes use of superbright LED has been developed. The maximum light power obtained from the developed system using a single superbright LED (TLRA 190P, λ
c=660nm) was approximately 1.5 watt at pulse width 24nsec and the pulse light intensity was stable over 30 hours, corresponding to 1×10
7 shots running at a pulse repetition frequency of 100Hz. The pulse light generator is especially suited for the observation of high speed moving objects by stroboscopic methods due to its high power light output, incoherence, pulse width and temporal stability. The stroboscopic method of imaging has been successfully applied e.g., to the microscopic investigation of the breakup of a liquid fuel jet discharging from a nozzle into the atmosphere chamber which has cylindrical space and to the observation of the droplets produced during this breakup process. The results show that the fuel flows have very distinctive shapes just after the start of injection.
View full abstract
-
Toshiharu SUGIE, Masafumi OKADA
1996 Volume 32 Issue 3 Pages
313-319
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
Recently, the importance of the joint design of identification and control has been recognized, and several controller design methods based on the iteration of identification and controller re-design are proposed. However, in these methods, we can not expect the convergence of the iterative procedure. Furthermore, although various powerful robust control methods have been developed remarkably, it is difficult to take advantage of them. Therefore, in this paper, we propose a new method of the joint design of closed-loop identification and control. In the propose method, the convergence of the iteration is considered in the controller design stage, and we can make use of the
H∞ control methodology. In addition, we evaluate its effectiveness via experiments using a positioning servo system with flexible joint.
View full abstract
-
Kang-Zhi LIU, Tsutomu MITA
1996 Volume 32 Issue 3 Pages
320-328
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
This paper gives the parameterization of all controllers achieving the comprehensive stability which is the stability required both in regulation problems and in servo problems of linear systems. The parameters in the parameterization of comprehensive stabilizing controllers are analyzed. Further, it is shown that subject to some conditions, the obtained comprehensive stabilizing controllers possess an Internal-Model-like structure. It is also proven that in such cases the condition of comprehensive stabilizability implicitly indicates that the plant does not have transmission zeros capable of canceling the poles of the Internal-Model, as was well understood.
View full abstract
-
Based on The Small Gain Theorem
Noboru SEBE
1996 Volume 32 Issue 3 Pages
329-335
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
This paper is concerned with the conditions for integrity. Integrity is the property that closed-loop systems remain stable in the presence of failures of sensors and/or actuators. This property is important for multivariable control systems.
In this paper, a condition for partial integrity against failures of both sensors and actuators is derived. This condition is described with the augmented system, that is the closed-loop system without the sensor or actuator-connections where failures may occur. This paper also clarifies the relation between the previously known sufficient condition, strictly positive realness, and the passivity theorem. Finally, this paper proposes a less conservative sufficient condition for integrity based on the small gain theorem.
View full abstract
-
Hansheng WU, Koichi MIZUKAMI
1996 Volume 32 Issue 3 Pages
336-344
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
The problem of robust stability of time-delay system with structured state space uncertainties is discussed. A method is presented whereby a new class of sufficient conditions for robust stability of such an uncertain time-delay system are derived. It is shown that the stability robustness of such a system can be ensured provided that the eigenvalues of nominal closed-loop system (i.e. the closed-loop system in the absence of uncertainty and delay) lie to the left of a vertical line in the complex plane which is determined by a norm involving the structure of the uncertainties (including those corresponding to delayed state) and the eigenvector matrix of nominal closed-loop system. It is also shown that for time-delay systems with structured state space uncertainties, the robust stability condition derived here is less conservative than those reported in the control literature.
In addition, we only assume that the time-varying delay is a bounded and continuous function, and the robust stability condition derived here is independent of the delay. Therefore, our results are also applicable to systems with perturbed time-delay. Finally, a numerical example is given to demonstrate the validity of our results.
View full abstract
-
Yasumichi HASEGAWA, Tsuyoshi MATSUO
1996 Volume 32 Issue 3 Pages
345-354
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
The realization problem of linear representation systems had been presented with the following realization theorem. Realization theorem [For any input response map (equivalently, any input/output map with causality), there exist at least two canonical linear representation systems which realize it. Moreover, Any two canonical linear representation systems with the same behavior are isomorphic.]
And the finite dimensional of the systems had been also discussed. Moreover, for a data obtained by multi-experiment, the partial realization problem of the systems had been discussed. Where any non-linear systems with on-off control, any homogeneous bilinear systems and any
K-Σ-automata are examples of linear representation systems.
In this paper, being based on the above results, we will discuss a problem of the real-time partial realization. The problem can be roughly described as the follows.
[For a given partial input response map by single experiment, find uniquely a minimal dimensional linear representation system which realizes it.]
Introducing invertible linear representation systems, we will obtain the real-time partial realization theorem.
In the field of automata, the identification problem of automata have been solved with introducing the notion of “strongly connected”, which is weaker than the notion “invertible”. Namely, “strongly connected” implies “invertible” and the converse does not hold.
We conclude that the result of this paper is the extension of the finite dimensional, constant, linear systems to a non-linear systems.
View full abstract
-
Tetsuhiro NAKAMA, Hayao MIYAGI, Tsutomu NAGADO, Katsumi YAMASHITA
1996 Volume 32 Issue 3 Pages
355-362
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
It is well known that the direct method of Lyapunov is the most suitable method for on-line transient stability estimation of power systems. Several techniques have been proposed for the construction of Lyapunov functions, for this purpose. The effectiveness of these Lyapunov functions for power systems, however, has been demonstrated only on digital computer simulations, without experimental results. In fact, to utilize the Lyapunov functions for on-line transient stability estimation of actual power systems, it is necessary to measure state values and to manage system noise, operating time for actuating mechanical circuit breakers and computing time of the Lyapunov function well. Moreover, stability boundary obtained by the Lyapunov function is more or less moderate because the Lyapunov method gives only a sufficient condition. Hence, constructing superior Lyapunov function that estimates stability boundary closer to true stability boundary is needed. In this paper, a computer-controlled on-line transient stability estimation system for power systems is constructred, taking these problems into account. An extended Lyapunov function is utilized to estimate the transient stability of the power system with GOV. The experimental results show that the proposed estimation system can act satisfactorily for estimating the transient stability of the actual power system.
View full abstract
-
Kazuo TANAKA, Takahiro KOSAKI
1996 Volume 32 Issue 3 Pages
363-368
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
This paper presents fuzzy backward movement control of a trailer type mobile robot via parallel distributed compensation. A Takagi-Sugeno type fuzzy model is used to represent the nonlinear dynamics of the trailer type mobile robot. The parallel distributed compensation achieves a fuzzy controller design intuitively and easily. In the parallel distributed compensation, each control rule of fuzzy controller can be designed from the corresponding rule of the Takagi-Sugeno type fuzzy model which represents the nonlinear dynamics of the trailer type mobile robot. Furthermore, we analyze the stability of the fuzzy control system via Lyapunov approach. A Lyapunov function is found by using a construction method in the stability analysis. The experimental results show that the designed fuzzy controller effectively realized backward movement control of the trailer type mobile robot so as to avoid jack-knife phenomenon.
View full abstract
-
Hiroaki GOMI, Mitsuo KAWATO
1996 Volume 32 Issue 3 Pages
369-378
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
The mechanical impedance of the human multi-joint arm during point-to-point movement in horizontal plane was investigated by using a high-performance manipulandum and an estimation algorithm newly developed. Ratios of the elbow, shoulder, and two-joint stiffness were greatly altered during movements. As a result, the shape and direction of a hand stiffness ellipse were changed, and the stiffness in a certain direction was sometimes lower than that of posture maintenance. Although the joint stuffiness during movements was increased, it was still lower than that previously assumed in the equilibrium-point control hypothesis. This revealed that the equilibrium trajectory (especially, in the velocity profile) was quite different from the minimum jerk trajectory, indicating that the minimum jerk trajectory could not be a reference trajectory sent to neuro-muscular servo controller based on the equilibrium-point control hypothesis. Rather it suggests that the central nervous system regulates muscle activations to produce appropriate torque patterns based on an internal model of arm dynamics, and to form the arm impedance adequately for interactions with environments.
View full abstract
-
Hajime NISHIDA, Yutaka MATSUMOTO, Yutaka YAMAMOTO
1996 Volume 32 Issue 3 Pages
379-388
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
This study presents a new method to prune hidden units in trained recurrent neural networks. Our approach is based on Moore's method for reduction of linear systems. To improve generalization ability and to reduce computational complexity, pruning methods of neural networks have been actively studied. However the choice of pruned neurons has been more or less empirical in previous methods. In the present method we first linearize the output function of neurons to approximate trained recurrent networks by linear systems. Then Moore's method, which removes the less controllable/observable subsystem after the original system is coordinately transformed to an internally balanced system, is applied to the resultant linear system for size reduction. Finally the reduced system is converted to a non-linear recurrent network for retraining. Numerical results show that our method can prune hidden units successfully with a small amount of retraining computation.
View full abstract
-
Moonhong BAEG, Hideki HASHIMOTO, Kiminori HASEGAWA, John B. MOORE, Fum ...
1996 Volume 32 Issue 3 Pages
389-395
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
A key problem in robotics is the estimation of the location and orientation of 3-D curved objects from surface measurement data. This is termed pose estimation. We assume that parts of 3-D curved objects are approximated to quadratic surfaces. Fundamental task is to estimate the pose of quadratic surface from measured data. A solutions for this task facilitates pose estimation for more complex object. Our pose estimation problem is converted to a non-linear optimization problem that minimize an error objective function between the measured surface data and the one of CAD model. The authors study gradient flows on the Lie groups towards a solution of the pose estimation problem of quadratic surfaces. In this paper, the projected gradient flow of the objective function onto the manifold
SO(3)×
R3 is derived and converses to an equilibrium point as usual steepest decent methods. Discretizations of the flow leads to recursive numerical methods for pose estimation. The proposed algorithm has been tested to show its validity using synthetic and real data.
View full abstract
-
Fumihiro OHBA, Hironao KAWASHIMA, Takao WADA, Haruyoshi MIZUNO
1996 Volume 32 Issue 3 Pages
396-400
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
The well used and the traditional methods in signal processing of heart rate variability (R-R interval) have been based on power spectrum analysis or linear modeling. But the essential properties of heart rate control system seems to lie in non-linearity and a method to grasp the characteristics caused by this non-linearity is needed.
From this point of view, correlational dimension analysis and bispectrum analysis to R-R interval in electrocardiogram were conducted. Non-linear properties of the time series is reflected in the trajectory depicted in certain order phase space. Correlational dimension is a geometric characteristics of the attractor in phase space. Correlational dimension analysis is a useful method to measure a fractal dimension of attractor in phase space from field data, and the spatial structure of the time series in particular order phase space can be obtained. Bispectrum analysis can be interpretated as a second order transform of third order correlation. This quantity can evaluate the interactions between frequency components which are caused by non-linearity of the system.
The data used in this analysis were taken from healthy control subjects who were under a mental stress condition (arithmetic calculation) and a non-stressed condition. By comparing these two kinds of data, the following result were obtained:
Under the non-stressed condition, R-R interval data shows a certain kind of spatial structure in a limited order phase space (correlational dimension was about 3 to 4), and it seems to have a non-linear mechanism in the source of the heart rate control system. Besides, under the stressed condition, this property was banished and R-R interval data did not have specific spatial structure in any order phase space, and similar to the property of the gaussian white noise.
View full abstract
-
Yuji SATO, Hiromitsu ISHII
1996 Volume 32 Issue 3 Pages
401-408
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
Recently, the avoidance from ship collisions becomes a mandatory problem in the viewpoint of maneuvering ships safety in crowded or crossing areas. Today, information for the steering is mainly limited to pilot's eye measurement and radar information.
As for radar information, the detecting the course direction or turning of a target vessel urgently and accurately with an accuracy is difficult. Moreover, information on the bigness and the altering course of a target vessel obtained by eye measurement is regarded as seriously important for the avoidance. Then, the authors have studied on the collision avoidance systems that make use of image information instead of eye measurement.
In this study, they obtain a method for measuring course of a target vessel and evaluation of the collision risk by the extracted image of target from IR image.
As the result of this study;
(1) The image of the vessel is extracted from IR image base on the threshold set method by using area information.
(2) The course and the length of a target vessel are obtained from the number of the pixels corresponding to horizontal length of the vessel. And the altering course of a target vessel is detected on real time.
(3) As for the direction of the prow of the target vessel, it is done from the specified centroid of the vessel image.
(4) The amount of the altering course necessary for avoidance is obtained from the meeting condition predicted from radar and the image. And effect is evaluated.
It is demonstrated and evaluated that the effective and simultaneous use of radar and the image can improve the performance of collision avoidance.
View full abstract
-
Motohide UMANO, Shiro FUKUNAKA, Itsuo HATONO, Hiroyuki TAMURA
1996 Volume 32 Issue 3 Pages
409-416
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
We extract fuzzy rules from data using a fuzzy neural network without a prior information. First, we set the number of initial fuzzy rules based on the number of training data. Next, we generate the specified number of initial fuzzy rules that have less wasteful membership functions using self-organization learning by T. Kohonen. Finally, we tune and prune fuzzy rules using the similar method to back-propagation learning with forgetting by M. Ishikawa. We apply the method to problems of the iris classification by R.A. Fisher and the diagnosis of electric transformer by gas in oil, and we have very good results.
View full abstract
-
Shuichi ADACHI, Tomoaki EDA
1996 Volume 32 Issue 3 Pages
417-419
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
A new grey-box modeling method via continuous-time model is proposed, where
a priori knowledge on physical parameters is expressed in a deterministic manner. It is shown that the proposed method is more effective for smaller data case through numerical examples.
View full abstract
-
Tomohiro KUBO, Etsujiro SHIMEMURA
1996 Volume 32 Issue 3 Pages
420-422
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
The vibration control of a cantilever beam with an active mass damper is considered. It is shown that a linear quadratic regulator can be constructed by a lumped-parameter-part control, without solving any infinite dimensional Riccati equation.
View full abstract
-
Isao YAMAGUCHII, Takashi KIDA, Tokio KASAI, Masaaki SANO, Keiji KOMATS ...
1996 Volume 32 Issue 3 Pages
423-425
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
This paper studies modal parameter identification experiments of flexible space structure as single-input multi-output system using accelerometer outputs. For the purpose, Eigensystem Realization Algorithm is modified to be applicable to the practical problems and its accuracy is enhanced by adding the iterative least square process.
View full abstract
-
Haruhiko KIMURA, Takehiko ABE, Hidetaka NANBO, Takashi OYABU
1996 Volume 32 Issue 3 Pages
426-428
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
In order to decrease the number of combinations in matching, we propose a new data structure of α-memory such that, as a rule is given, it can immediately point out working memory elements satisfying the rule's condition element.
View full abstract
-
Shigemasa TAKAI, Toshimitsu USHIO, Shinzo KODAMA
1996 Volume 32 Issue 3 Pages
429-431
Published: March 31, 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS
In this paper, we derive a necessary and sufficient condition under which there exists a command-based supervisor for a given closed language under partial observations. The result generalizes the one obtained in the Ramadge-Wonham framework.
View full abstract
-
1996 Volume 32 Issue 3 Pages
432
Published: 1996
Released on J-STAGE: March 27, 2009
JOURNAL
FREE ACCESS