All issues
Volume 36, Issue 1
Displaying 1-17 of 17 articles from this issue
- |<
- <
- 1
- >
- >|
-
Masahiro ISHIBASHI, Masaki TAKAMOTO2000 Volume 36 Issue 1 Pages 1-9
Published: January 30, 2000
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSDischarge coefficients of 10 super-accurate nozzles whose machining error were smaller than ±1μm were measured using a constant volume tank system at absolute pressures of 100 to 500kPa with the Reynolds number of 4×104 to 1.2×106. The measured discharge coefficients agreed within ±0.1% with a reference curve, which had been measured using other 24 super-accurate critical nozzles at the atmospheric pressure or lower. Discharge coefficients of one nozzle among the 10 showed a sudden decrease by 0.1% at the Reynolds numbers around 106, which was caused by a boundary layer transition and defined the valid maximum Reynolds number of the reference curve. All of the equations and constants needed to calculate a flow rate using the reference curve are listed. An uncertainty analysis of the constant volume tank system, developed as the national standard and used to calibrate the nozzles, is described.View full abstractDownload PDF (2343K) -
Yoshiya TERAO, Masaki TAKAMOTO2000 Volume 36 Issue 1 Pages 10-15
Published: January 30, 2000
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe large water flow calibration facility, the primary standard of water flow rate in Japan, is one of the largest calibration facilities for water flow meters in the world. This facility covers the flow rate range of 0.014 to 0.83m3/s (50 to 3000m3/h). Flow rates are maintained by seven pumps and an overflow head tank. A water flow meter under test is installed in one of the test lines whose inner diameter (D) ranges from 100 to 1100mm with a minimum upstream straight pipe length of 100D. The flow rate through the flow meter is measured using the gravimetric method. The uncertainty of this facility has been estimated experimentally and analytically in detail in accordance with the ‘ISO Guide to the expression of uncertainty in measurement.’ As a final result, the expanded uncertainty of the flow rate (k=2) was estimated to be 0.034%. The major error sources and their contribution to the standard uncertainty of the flow rate were the measurements of (a) the mass of the water in the weighing tank (0.0130%), (b) the time duration for which the water flow was diverted to the weighing tank (0.0085%), (c) the buoyancy for correction of the mass of the water in the tank (0.0070%) and (d) the density of the water (0.0025%). The uncertainty of the mass measurement of the water in the tank, which is the largest error source, was mainly due to the limits in the reproducibility of the weighing scale. From this result, an improvement of the calibration method of the weighing scale was proposed. It was clearly shown that this kind of uncertainty analysis is very important in assessing and improving the performance of a water flow calibration facility.View full abstractDownload PDF (2329K) -
Akira HIRABAYASHI, Hidemitsu OGAWA, Tatsuya MIZUTANI, Ken NAGAI, Katsu ...2000 Volume 36 Issue 1 Pages 16-25
Published: January 30, 2000
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSWe devise a new method of increasing the data acquisition and processing speed in a scanning white-light interferometer for surface topography measurement. We introduce the concept of a characteristic function which is smoother than interference fringe (IF) and has a maximum at the same point as IF. We also show that IFs are band-pass signals, instead of low-pass signals. By using the sampling theorem for band-pass singals, an interpolation formula for the characteristic function is derived, which leads us to a fast surface profiling method. The number of sample points are reduced to 1/24 of that for the conventional method. Experimental results on a 9.947μm step height standard show that our method achieves the same level of accuracy as the conventional one.View full abstractDownload PDF (3158K) -
Ryosuke HOSAKA2000 Volume 36 Issue 1 Pages 26-31
Published: January 30, 2000
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSA new method to measure eye-movement is developed. The method is named capacito-oculography (COG). In this method, capacitance change between surface of an eyeball and an electrode is used for measurement. Pair electrodes are set up to front of the eyeball. A distance between both surface change when eye movement occurs. The change of the distance becomes the capacitance change of the space. The change also equivalent to a change in capacitance contained in a tank circuit of an oscillator. Therefore, a carrier wave is modulated by the capacitance change based on the eye movement like as a frequency modulation (FM). An information of eye movement can be measured by discrimination of the FM modulated signal.View full abstractDownload PDF (2172K) -
Kazushige MAGATANI, Hirohito IWAYA, Hironori YAMASAKI, Kenji YANASHIMA2000 Volume 36 Issue 1 Pages 32-37
Published: January 30, 2000
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSOur final goal is the development of the auto-navigation system which supports activities of the visually impaired without help of others. This system measures a current position of the visually impaired user, calculates the route from this position to the destination, and guides the user to the destination by using artificial voice. In this system, DGPS technique is mainly used in order to measure a position of the user. However, DGPS does not work in the place where the system cannot receive the information from GPS satellites. In such place, our system estimates a position of the user by using the developed new method (self-positioning method). An acceleration sensor, a terrestrial magnetism sensor and a gyroscope which are equipped on a user are used in order to track the moving of the user. The number of steps is measured by an acceleration sensor, and the direction of progress is estimated by a terrestrial magnetism sensor and a gyroscope in our self-positioning method. Subject with perfect vision was tested with our new tracking method. From this experiment, this method was shown to be a very valuable one to track a position of the people without GPS.View full abstractDownload PDF (3195K) -
Shigenori SANO, Kazuya OGATA, Yoshikazu HAYAKAWA2000 Volume 36 Issue 1 Pages 38-45
Published: January 30, 2000
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSWhen a control system is designed, it is important to make a model which can explain the input-output relation precisely. To identify the model's parameter as precisely as possible, it is important to make the experimental signal sufficiently informative and less affected by the disturbance or the unstructured model uncertainty. In this report, the upper bound of the estimated parameter error is proposed when the unstructured model uncertainty exists. This upper bound can be used to determine the optimal identification input.View full abstractDownload PDF (1974K) -
Tomoki MIYAZATO, Tong ZHOU, Shinji HARA2000 Volume 36 Issue 1 Pages 46-52
Published: January 30, 2000
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSWe introduce a probabilistic measure named Model Set Unfalsified Probability (MSUP) for model set validation. We derive an upper bound and a lower bound of MSUP and show that the lower bound computation can be reduced to a convex optimization. Simulation results confirm that the probabilistic approach more appropriately evaluates the suitability of a model set in robust controller design than deterministic approaches.View full abstractDownload PDF (1703K) -
Hiromi MOCHIYAMA2000 Volume 36 Issue 1 Pages 53-59
Published: January 30, 2000
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSA desirable kinematic structure of a manipulator with hyper-many degrees of freedom is discussed from a theoretical viewpoint of shape control. Conditions for the kinematic structure are derived from a geometrically natural requirement that its kinematic equations tend to Frenet-Serret formulae as its number of degrees of freedom increases. The obtained result indicates that the manipulator should have two-degree-of-freedom joints each of which consists of a twist and one-sided bend.View full abstractDownload PDF (2714K) -
Hidehiko SUGIMOTO, Kazuo WASHIDA2000 Volume 36 Issue 1 Pages 60-67
Published: January 30, 2000
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe modified repetitive control system is designed using a modified repetitive controller with a low-pass filter incorporated into the positive feedback loop for the time-delay element. The modified repetitive controller, however, functions only for the gain determined by the low-pass filter at the basic frequency as the desired value and the frequency obtained by multiplying it by integers. This problem can be solved by two methods: one is by using a modified repetitive controller with linear phase filter and the other by using a modified repetitive controller with corrected dead time.
This paper describes the results of our study about the above two methods to identify the differences between them in their control characteristics and formulate the guidelines for their use as summarized below:
(1) The modified repetitive controller with linear phase filter can basically function as well as the modified repetitive controller with corrected dead time only if its filter is of higher order. However, if the spectrum of the desired value is distributed over a wide frequency range, the use of a higher-order filter in the controller to increase its gain conversely results in narrowed frequency bandwidth, not providing so much improvement in its control performance as expected, and frequently presenting the practical problem of increasing the time required for the computation. Notwithstanding this, if the spectrum is locally distributed in a narrowed frequency range close to the basic frequency, an increase in the order of the filter used in the controller as allowed by the computation time enables its control performance to be improved.
(2) The modified repetitive controller with corrected dead time, even if its filter is of first order, can perform its control function comparable to that of the modified repetitive controller with higher-order linear phase filter. This controller is a practical control method, particularly having the great advantage of increasing its gain to∞for direct current.View full abstractDownload PDF (1927K) -
Masatoshi NAKAMURA, Satoru GOTO, Nobuhiro KYURA2000 Volume 36 Issue 1 Pages 68-74
Published: January 30, 2000
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSHigh accuracy performance of contour control is required for industrial robot arms such as sealing process, laser cutting and so on. This paper describes contour control of articulated robot arms by introducing nonlinear separation control concept. The proposed method consists of nonlinear static compensation of the inverse kinematics and kinematics, and linear dynamics compensation by using the modified taught data method in the joint coordinates. The effectiveness of the proposed method was assured by simulation results and experimental results using an actual industrial robot arm.View full abstractDownload PDF (1458K) -
Yasue MITSUKURA, Toru YAMAMOTO, Masahiro KANEDA2000 Volume 36 Issue 1 Pages 75-81
Published: January 30, 2000
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSPID control schemes have been widely used in most of process control systems represented by chemical processes for a long time. However, it is a very important problem how to choose the optimal control parameters, i.e., PID parameters, because these parameters give a great influence on the stability and the control performance. On the other hand, self-tuning PID control has proposed for systems with unknown parameters. However, to the best of our knowledge, there is few scheme which is effective for the system with time-varying parameters.
In this paper, a new self-tuning PID control scheme is proposed, which is able to deal with such a system. The proposed scheme is derived based on the relationship between PID control and generalized minimum variance control (GMVC) laws. Furthermore, a suitable set of some user-specified parameters included in the GMVC criterion is sought by using a genetic algorithm recursively.View full abstractDownload PDF (1226K) -
Kazunori MOCHIDUKI, Takashi KIDA, Keiji KOMATSU, Isao YAMAGUCHI2000 Volume 36 Issue 1 Pages 82-89
Published: January 30, 2000
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper studies a decentralized control of space structure which is constructed in orbit by connecting a new module to an already existing structure. For such an expanding system, it is practical to design only the local controller of the new module so as to stabilize the system at the connecting instance, without changing the existing system. To this end, we propose a design procedure in μ synthesis framework by using a descriptor equation as well as a state equation. Then, their efficiencies are illustrated and compared by a numerical example where two flexible modules are connected through a flexible joint which has uncertain parameters.View full abstractDownload PDF (1644K) -
Shinzo KITAMURA, Yuzuru KAKUDA, Hajime MURAO, Jun GOTOH, Masaya KOYABU2000 Volume 36 Issue 1 Pages 90-97
Published: January 30, 2000
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe design process of artifacts is first formulated based on the concepts of structure, property and function, then a new mathematical definition of specification is given. Under this formulation, the design problem is defined as an inverse of the mapping from the structure set to the specification set. Next, the design procedure of human designers is considered to be composed of an iteration of the processes of a new structure generation, computations (or experiments) of its function through the forward mapping and the comparison of this function with given specification. The role of this structure generation and comparison is interpreted as an emergent property in the design problem formulated above, and this role is relpaced by emergent computings. Finally, genetic algorithms are applied for two examples; the first one is a synthesis of passive linear filters using inductors and capacitors as elements, and the second is a synthesis of moving multi-link robots using rotating joints and links as elements. These two trials show an emergence of desiable structures and function adapting to a specification and an environment.View full abstractDownload PDF (3993K) -
Baiquan LU, Junichi MURATA, Kotaro HIRASAWA, Jinglu HU2000 Volume 36 Issue 1 Pages 98-107
Published: January 30, 2000
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn this paper, we present a new learning method using prior information for three-layered neural networks. Usually when neural networks are used for identification of systems, all of their weights are trained independently, without considering their inter-relation of weights values. Thus the training results are not usually good. The reason for this is that each parameter has its influence on others during the learning. To overcome this problem, first, we give exact mathematical equation that describes the relation between weight values given a set of data conveying prior information. Then we present a new learning method that trains a part of the weights and calculates the others by using these exact mathematical equations. This method in almost all cases keeps a priori given mathematical structure exactly during the learning. Numerical computer simulation results are provided to support the present approaches.View full abstractDownload PDF (2073K) -
Hiroyuki OKAMURA, Tadashi DOHI, Shunji OSAKI2000 Volume 36 Issue 1 Pages 108-115
Published: January 30, 2000
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn recent years, auto-sleep functions for hard disk drives of computer systems have been paid much attention in term of the power saving. In this paper, we suggest the optimal design for sleep mode under the performance criterion called power effectiveness. In particular, we consider the situation where access requirements are accumulated in the buffer of a system. When the access requirements occur in accordance with the homogeneous Poisson process, the optimal sleep time which maximizes the power effectiveness is analytically derived. For more general cases, two kinds of approximate methods are proposed.View full abstractDownload PDF (1925K) -
Shigenobu ITOH, Satsuki NAKAJIMA, Ryoichi HARADA, Ryoji KAWATANI2000 Volume 36 Issue 1 Pages 116-118
Published: January 30, 2000
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe design of reduced-order controllers in the set of LSDP-controllers is proposed in this paper. In order to reduce the order of those controllers with a free parameter, Φ, two equations need to be solved: For solving the equations, we adopt the controllability canonical form by Luenberger as Φ. For m-outputs (MIMO) system, we can reduce N-th order LSDP-controller to (N-m)-th one. Several constants in the Φ are determined by Genetic Algorithms. We compared the performance of controllers and found the results the original controller and the reduced one to be identical.View full abstractDownload PDF (405K) -
Tielong SHEN, Katsutoshi TAMURA2000 Volume 36 Issue 1 Pages 119-121
Published: January 30, 2000
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSRobust exponentially passivation via output feedback is addressed for a class of uncertain nonlinear systems. The uncertainty is described by gain bounded unknown functions on the state. A nessecary and sufficient condition is derived for the uncertain systems being robust exponetially passive respect to C1 storage function. Using this condition, a semi-global output feedback controller is proposed which renders the uncertain nonlinear system robust exponentially passive.View full abstractDownload PDF (437K)
- |<
- <
- 1
- >
- >|