In this paper, we propose two sub-optimal control methods of sampled-data systems, which consider the trade-off between performance and the number of samplings. They are as follows: 1) Multirate sampled-data control method, in which each component of input may have a different constant sampling interval. 2) Quasivariable interval sampled-data control method, in which the sampling interval of each component of input need not be constant and can have integer multiple values of a basic interval T.
Two simple numerical examples show these methods are effective for practical uses.

抄録全体を表示

This paper considers the model following servo system synthesis method in which optimization techniques like Kreindler are used to design controllers that make the step response of the plant should be kept close to a specified ideal step response of the model.
The controller is named model following servo and has several useful characteristics as follows.
(1) The controller includes the servo controller to exclude the offset in the presence of the constant disturbances or the parameter perturbations of the plant.
(2) The controller for constant reference signals and constant disturbances is treated in this paper, but its generalization to more general reference signals and disturbances can be easily done.
(3) The performance of the controller is as good as that of the adaptive model following controllers.
(4) The structure of the controller is simpler than that of the adaptive ones.
Two examples demonstrated the above mentioned characteristics.

抄録全体を表示

This paper is concerned with a new approach to estimating probability density function when several low-order moments or cumulants are known. The problem is to find an algorithm efficient for estimating probability density functions under the policy to maximize the entropy involved. In deriving this technique low-order moments, which are beforehand determined by any conventional procedures, are put as constraints. Estimation is then reduced to a problem to solve a system of nonlinear equations, whose resolution is performed by the Newton-Raphson method. The main features of this procedure are as follows: Estimation can be made using a rather simple system of equations and is significantly effective in applying to non-Gaussian distributions. Finally, as an example, this method is applied to reactor noise data and results obtained are in good agreement with those by histogram.

抄録全体を表示

We propose a new quantitative method for systems analysis that enables investigation of trends of people's needs by analyzing opinion surveys. We have applied the method to a nationwide opinion survey in Japan which was conducted in 1978 to obtain people's assessments of the degree of satisfaction and importance on 60 concerns of quality of life. Various interesting findings obtained by our quantitative method are presented, many of which support observations and qualitative analysis on the directions of people's needs for well-being in Japan in the late 1970s.
We first define an indicator that expresses the degree of a change of people's interest when their interest moves from one need to another after the former need is statiated. We present a hypothesis by utilizing the indicator, and obtain a hierarchical structure of people's needs based on the hypothesis. Finally, we construct a structural model that illustrates trends of people's needs for the improvement of quality of life, and we describe findings elucidated by the model.

抄録全体を表示

This paper is concerned with the design of a discrete time adaptive observer having exponentially converging properties to estimate state variables and identify parameters of an unknown linear time-invariant single-input single-output discrete time system. The observer structure is based on Kreisselmeier type parametrized representation of the system, but in this study parameter adjusting law is obtained by a new scheme. The adjusting law is determined by equations forcing an evaluation of errors into decreasing exponentially. That is, the adaptive observer is designed with an exponential rate of convergence in respect to parameter mismatches. The convergence rate is related on small number of design parameters of the observer and a high rate of convergence can be obtained. A sufficient condition for this property is obtained. An example of the thirdorder system is given to show the effectiveness of this design scheme.

抄録全体を表示

This paper defines a condition of well-posendness for general linear systems and investigates its meaning. Originally, the condition was prorposed for basic linear systems such that a linear feedback is applicable to them. This paper claims that the condition is as basic a condition as causality and stationarity.
Let S⊂X×Y be a linear system and ρ₂₀: X→Y its input response. Let L={K|K: Y→X} be the class of static linear feedbacks of S. Then (^∀K∈L)((I-Kρ₂₀): D(S)≅D(S^K)) is called the well-posedness condition where S^K is generated from S by a linear feedback K.
This paper shows that if the well-posedness condition is satisfied, (i) a state space representation of S^K is explicitly determined by that of S, and (ii) the input response and the state response of a state space representation satisfy special properties.
One of the purposes of this study is to classify the equivalence of controllability and the mode controllability in the system theoretic sense. The results of this paper strongly suggests that the equivalence may be explained by the combination of causality, stationarity, finite dimensionality of the system core and the well-posedness condition.

抄録全体を表示

A new method aimed at non-contact velocity measurement of moving objectsis proposed. In this method, ultrasonic wave is emitted, the echo field reflected by moving objects is detected by receiver array, and recorded for a certain period of time. The velocity vector field is reconstructed numerically from the recorded signal using digital beamforming techniques. Whole distribution of velocity vector components, both the longitudinal and transverse components can be measured simultaneously. The possible applications are non-invasive measurement of blood flow, observation of high temperature gas, or investigation of moving powder, etc.
In this paper, 2-dimensional case is discussed for feasibility study, however, the principle can be extended to 3-dimensional measurement. The measurement system emits monochromatic sinusoidal ultrasonic wave through a single transmitter and it detects the echo by a linear array of receivers.
Two velocity estimating algorithms are proposed. The accuracy of velocity measurement by the each algorithm is evaluated as the square root of the minimum obtainable variance of the estimate. Several fundamental relations for selection of parameters are obtained.
The experimental measuring system is built utilizing ultrasonic wave in air (40kHz). Computer simulation and experimental results prove the feasibility of the method.

抄録全体を表示

A fixed point blackbody furnace is an effective means for Calibrating radiation thermometers. It is not commonly used because of its complicate structure and troublesome operation. This paper deals with a set of fixed point blackbody furnaces which are simple in design and easy to operate. Metals for the fixed points are Cu, Ag, Al, Sb, and Zn. Their temperatures range from 1085°C to 420°C.
The graphite crucible in a stainless steel capsule contains about 26cm³ metal of 0.99999 nominal purity and has a cylindrical cavity of 10mm diameter and about 46mm length. The emissivity of the cavity is calculated to be 0.999±0.0005 assuming that the surface of the graphite is diffuse and its intrinsic emissivity is 0.85 at 900nm, and considering that the aperture made of stainless steel is 6mm in diameter.
Twelve trenches, cut in the length direction on the outer surface of the stainless steel capsule in an alumina tube, hold a nichrome heater winding. The furnace is a cylinder of 10cm diameter, 25cm length and 3kg weight, and consumes so little energy that input electric power of 400W raises the temperature of the capsule up to 1100°C in about an hour and a half, and that of 240W keeps the capsule at this temperature. Argon gas is blown into the capsule so as the crucible is not oxidized.
Freezing curves are measured by a 900nm silicon narrow band radiation thermometer. Plateau durations are longer than 10 minutes. The errors of the temperatures of the cavities in the realization of the fixed points are estimated to be -0.1±0.1K, and the accuracies of the fixed points for the calibration of the silicon radiation thermometer to be 0.3K.

抄録全体を表示

In Manufacturing technology of LSI, exposure, etching and other processes are iterated using several kinds of photomasks in order to form electrical circuit pattern on the silicon wafer. Therefore mask alignment is indispensable.
We tried to detect the translational displacements of IC utilizing the laser holography technique. Then real time on-line positioning control by a minicomputer was performed. The laser beam is divided into two beams, one of which is projected to a model IC and the reflected light from the model IC is projected to a photographic plate after passing through a Fourier transform lens. The other beam is directly projected to the photographic plate after passing through a beam expander, and after development the hologram is obtained.
Next, the reflected beam by the IC chip to be positioned which has the same pattern as that in case of hologram making is porojected to the hologram, and the diffracted light from the hologram generates an optical spot which is an autocorrelation function of the IC pattern. As movement of the spot is proportional to the displacement of the IC chip, measurement of the displacement of the IC chip can be done by detecting the spot using a photoelectric microscope which has L-shaped slits. The measurement data are sent from the positioning pulse generator to a minicomputer, thus positioning control is done in accordande with its command.
The positioning accuracy obtained is ±0.18μm (confidence coefficient: 99.7%) which is monitored by a laser interferometer.
This method has the feature that any marking on the IC chip is not necessary and that positioning of the IC chip is done using its own pattern.

抄録全体を表示

A rectification algorithm for geometric correction of images obtained by airborne scanner is described. It is based on parameter estimation of a polynomial model of distortion by ground control points. Computer simulation described here shows that estimation of the approximate order of distortion is useful for decreasing residual errors resulting after correction of images especially when ground control points are not uniformly distributed in observed images. PSS (prediction sum of squares) and MSEP (mean square error of prediction) were used for the estimation. They showed almost the same validity. We can estimate the order of distortion automatically by them. Evaluation of residual errors on selected test points in an image resulting after correction with order estimation was implemented. The averrage precision of the correction for an image with 400×400 pixels was achieved at a level from about 2 to 4 pixels.

抄録全体を表示

A high sensitivity (the ratio of the output pressure increment to the increment of the gap D between the nozzle and the measured object) of a back pressure and high supply pressure type of conventional pneumatic micrometer is accompanied by a small D and a large contraction ratio (A_e/A₁, A₁: section area of inlet, A_e: section area of exit of duct). On the other hand with small contraction ratio, the sensitivity lowers, though a large D can be expected.
While decreasing its velocity, a supersonic channel flow interacts on the boundary layer adherent to the wall resulting in a pseudo-shock which consists of a series of shocks and subsonic mixing region. The position of pseudo-shock is sensitive to the pressure ratio Φ (the ratio of supply pressure p_os to back pressure p_b).
In this paper, the static characteristics of two sorts of supersonic air gap sensors (one being of mono throat and the other of double throat) are described which have been manufactured for trial and can gain a large output pressure through shocks. They have comparatively large D (1.4∼1.6mm) and a high sensitivity (S_pmax=1.94kPa/μm for mono throat sensor and 11.00kPa/μm for double throat sensor).

抄録全体を表示

It is necessary to know the conditions in which sound affects an attached water jet and clarify these phenomena for practical application of acoustically controlled fluidic devices in oceanographic developments.
In the present papar, the effects of applied sound on the attachment of water jets to a curved wall are experimentally investigated from the changes of the deflection angle of jets and jet centerline on the basis of the fundamental characteristics of two-dimensional water jets sensitive to sound. In addition, the transient response of the test device with a receiver located in the jets is examined by using sound as a control input in order to discuss the possibility of applying the phenomena of the attached jets to fluidics. And the attachment phenomena of water jets affected by sound are considered from the statistical inference of the response time.
As a result, the following facts are especially confirmed:
1) Three different phenomena of wall attachment of a water jet sensitive to sound are fundamentally recognized. It is especially worth noting among them that an attached jet separates from the wall by sound and changes into a separated jet.
2) The dynamic behaviours of an attached water jet affected by sound are related to the formation of low pressure bubble region, the irregularity of fluctuations in the jets, and the stability of attachment.

抄録全体を表示

Recently, the positioning control using a DC servomotor is a fundamental technique which is common to many mechanisms, including data communication I/O equipment such as a serial printer, a magnetic disk memory, etc. Therefore, it is necessary to be faster and more accurate.
In order to speed up the point-to-point positioning control, it is necessary to consider the oscillatory characteristic of the positioning mechanism, and it is important to minimize the input energy for positioning.
For the former, the positioning control problem is formulated as an optimal regulator problem with conditions of resting the oscillatory load of one degree of freedom at the decided position and time, and of minimizing the input energy. It is concretely shown that, by the time-varying feedback system, a high speed positioning control can be realized, whose positioning time T is nearly equal to the natural oscillation period T₀ of the mechanism,
Moreover for the latter, the natural oscillation influence on the minimum-energy positioning control characteristics was investigated, Results, arranged in ratio λ of T to T₀, are as follow.
(1) Input energy P for an oscillatory load is greater than or equal to input energy P₀ for a rigid load.
(2) In a comparatively slow region, such that λ>λ₁≈1.43, P and maximum nondimensional current |u|_max for the oscillatory load are almost equal to those for a rigid load.
(3) In a fast region, such that λ<λ₁≈1.43 as λ decreases, P and |u|_max for the oscillatory load increase more rapidly than those for a rigid load.
(4) Maximum distortion |x₁-x₃|_max increases rapidly as λ decreases.

抄録全体を表示

In controlling the articulated manipulator in real time, various kinds of operations and computations should be processed as soon as possible in order to attain dynamical movements in good quality. This paper proposes an approximation method which uses three order polynomial functions to calculate each angle value of a manipulator for a short time. In deciding the coefficients of these polynomial functions, two methods are used. One is the method of least square using actual data which are calculated accurately by solving the nonlinear equations, and another is the training method. The computer simulation shows a good performance of this method.

抄録全体を表示