A new technique for the measurement of particles in powder processing has been developed to identify the shape, size and position of multiform particles simultaneously and instantaneously. A new Multiplexed Matched Spatial Filter which is a single exposure hologram was employed for the new technique. In the identification tests, the input objects were discriminated by their autocorrelation peaks and their positions were acculately determined. It was concluded from the theoretical and experimental studies that the present method could be applied to simultaneously and instantaneously identify the shape, size, number and position of multiform particles in powder processing.

抄録全体を表示

Three dimensional distribution of radiation emitters is measured by a simple method using a random-coded aperture and a 2-dimensional (area) sensor. A radiation emitter casts the shade of the aperture onto the area sensor. The area sensor yields 2-dimensional image data. The depth from the aperture and the shift of an emitter in a plane parallel to the aperture modulate the magnitude and the shift of the shade, respectively. Demodulation from the output (image data) of the area sensor is conducted for every plane at some incremental distance from the aperture by cross-correlating the image data with the coded aperture pattern magnified according to the distance. The aperture was coded by a pseudo-random pattern (M-array composed of maximum-length sequence), whose auto-correlation function is approximately a Dirac's delta function. A maximum of the cross-correlation function indicates the existence of a radiation emitter. The procedure of cross-correlating with the magnified M-array is repeated for a number of planes at different distance from the aperture. Thus 3-D distribution of radiation emitter was measured.
The spatial resolution in the plane parallel to the aperture was about 5mm, and that in the direction of depth was about 50mm (at the depth of 300mm) for an object with 70×70×140mm³. The M-array we used was composed of 8-tuple m-sequence (period is 255), and the dimension of the aperture made of a 3mm thick lead plate was about 200×180mm² which includes 4×4 periods of M-arrays. The hole has 2mm diameter, and the pitch of the holes was 3mm.

抄録全体を表示

A calculation method of propagation characteristics and the cepstrum analysis of reflected waves are discussed for monopulse wave traveling perpendicularly to the multi-layered dielectric structure. The aim of this research is to detect the position of the air layer by processing the received underground radar signal. Using the cepstrum method, the received radar signal is decomposed into elementary reflected waves, when radar signal is composed of many waves passing through different paths due to reflection of transmission which occur at boundaries around the target.
By uses of the delay time of the elementary reflected waves, the position of the target hidden in invisible medium is estimated. The effects of the bandwidth of the antenna and random noise at the received signal on the cepstrum are discussed. An example of application of the cepstrum method to the experimental data of concrete, air layer and wet soil is also shown.

抄録全体を表示

Leakage current indicators using ECD (Electric Chromic Display) are developed. These indicators are characterized by the following points.
(1) Coloration density of the indicator shows electric charge which is integration of leakage current.
(2) The coloration density is held even if input current is stopped.
Fundamental characteristics of ECD are measured. Methods which hold coloring density, compensate coloring density depending on temperature and set the threshold of coloration are described. Leakage and ground fault indicators are fabricated. Minimum charges which display in leakage and ground fault indicators are 4.2±0.8mC, 4.9±1.1C, respectively.
It was confirmed that the developed indicators were effective in electric power field.
Leakage current indicators help to determine the equipment in which insulation has broken down from lightning etc. and are useful in finding the equipment in which insulation is deteriorating.

抄録全体を表示

Measure of concentration based on conductivity meter is necessary to investigate the kind. It is desired to indicate the concentration and kind of solution at the same time. Then new measurement method for electrolytic solution was studied. The feature of fundamental concept of the method of this measurement is to record Q-variation of the printed coil sensor continuously by a Q-meter during trasmission and diffusion of a solution into pure water through the cellophane transmission membrane. Three items of the method were studied. Improvements were made on measurement cell for removing the air bubble and on model formula for regression of Q-variation. And the results of principal component analysis for regression parameters were applied to identification. Four kinds, four concentrations of electrolytic selected on solutions were measured and analyzed. As the result of principal component analysis, it was found that two principal components represented original parameters. Therefore the efficiency of principal component analysis for this measurement and possibility of identification were indicated.

抄録全体を表示

This paper describes how to implement volume measurement effectively using the pneumatic method for automobile fuel gaging.
The pneumatic method estimates the remaining fuel in a closed tank by subtracting the gas volume in the tank from the tank capacity. The gas volume is calculated by applying the ideal gas law to the relation between the gas volume and the pressure in the tank. The gas volume is inversely proportional to the amplitude of the pressure change due to a small change in tank capacity.
In implementing this method, we faced the following basic problems:
(1) Bias error due to the flexibility of the fuel tank.
(2) Noise due to tank wall vibration and liquid sloshing.
(3) The measurement system must be compact and simple.
Problem (1) and (2) are substantial causes of error and noise in using this method and problem (3) is a specified prerequisite for automobile equipment.
We investigate how the flexibility of the tank influence the bias error and show the strategy used to solve it. We present a new compact sensing system composed of a loud speaker and two microphones and investigate the robustness of the system for noise in the tank when driving. Experiments using a real 55 liter tank verified the validity of the gaging system. The maximum error when driving on a very rough road was found to be±1.5 liters.

抄録全体を表示

Monocytes in charge of cellular immunity have the phagocytic function. It is extremely important to measure the phagocytic activity in the field of both basic and clinical medicine. But it is complicated and difficult to separate only monocytes from peripheral blood, and so there are no decisive methods of the measurement under the present conditions.
Therefore authors developed a new method which measured the phagocytic capacity of monocytes, using the mononuclear cells (mixture of monocytes and lymphocytes) in peripheral blood and the microscopic magnetic beads. This method is based on the principle, which attracts the monocytes phagocytized magnetic beads by magnetic force in a magnetophoretic bath effecting the direct magnetic field of the permanent magnet, and which separates the phagocytizing monocytes from the non phagocytizing monocytes and lymphocytes. This method has several advantages, such as, it is a simple, easy and quantitative procedure, compared with those currently in use.
This paper explaines the method of mesurement and the experiment evaluates the actual clinical data, which demonstrates the usefulness of the method.

抄録全体を表示

In this paper the concept of the negative margin of an algebraic Riccati inequality is introduced and its properties are discussed. Using these properties, H_∞ robust sub-optimal control problems are investigated for the plant with time varying structural uncertainties. Such a sufficient condition is given that a nominal H_∞ sub-optimal system has robust sub-optimality for structural uncertainties of the plant. When nominal parameters of the plant are given, a necessary and sufficient condition for existence of H_∞ robust sub-optimal controller by state feedback is derived and a design method is given.

抄録全体を表示

This paper is to propose a design method of MRACS (Model Reference Adaptive Control System) for unknown plants with slowly-varying disturbance w(t), by which the output error from the desired behavior can be reduced to a small value.
The control system is composed by using time delay controllers which compensate the effect of disturbance. To insure the stability of the control system an adaptive adjustment law with dead-zone is introduced. As analyzed theoretically, by the proposed control method the influence of the disturbance on the output will be reduced to the level of [w(t)-w(t-T)], where T is the delay time. So when the disturbance varies solowly, the steady state error can be made sufficiently small by selecting a small delay time T. Especially when the disturbance is constant, such as step-wise, it is possible to let the steady state error be zero. The effectiveness of the proposed method is illustrated by some computer simulations.

抄録全体を表示

Self-tuning control methods have been developed for the system with unknown parameters up to now. Furthermore, the rapid progress of the digital computers enables us to control a number of real plants based on digital control strategies. However, PID and I-PD control methods have been used for the real plants since the control parameters are specified by the proportional, integral and differential actions. It is easily to understand the physical meaning of the control system by these PID gains. Therefore, it is necessary to consider control methods that the control parameters are specified by the PID gains only.
In this paper, we construct a multivariable self-tuning controller with two-degree-of-freedom PID structure for the discrete-time systems. First, we construct a two-degree-of-freedom PID controller based on the model matching method. Next, we consider a self-tuning method of PID gain matrices. Many control strategies have been reported for auto-tuning and self-tuning PID controllers. However, there are not any design methods for the multivariable systems. Furthermore, we propose a design method of desired reference model matrices since the control performance depends on it. Finally, we illustrate a numerical simulation result to show the effectiveness of the proposed control algorithm.

抄録全体を表示

Motivated by the observation that if det(B)≠0 in a linear system (A, B), the controller design problem is trivial in some sense, the concept of full-controlled system will be introduced. It is shown every strongly attainable nonlinear control system can be decomposed to a series of full-controlled subsystems by introducing the virtual input variables. The decomposition is called the virtual decomposition. A nonlinear full-controlled subsystems do not have so good properties as linear ones. To discuss the stabilization problem of the virtually decomposed system, the concept of the practical stabilization has been introduced. A system is called practically stabilizable if the trajectory of the system has been confined in the arbitrary small neighborhood of the equlibrium point by the state feedback. A full controlled system is called completely decomposable if it admits the decomposition to scalar subsystems which are linear in input. It is shown that the completely decomposable full-controlled systems are practically stabilizable. Finally, a controller design method will be proposed for the virtually decomposed nonlinear systems.

抄録全体を表示

Recently, vibration control of a wheeled vehicle running over an irregular terrain has been studied. In order to increase the mobility of the vehicle and stability of circumstance sensor, it is very important to decrease the vertical, rolling and pitching vibration, which is caused by irregular terrain. Until now, however, almost all the studies have been carried out for the vertical vibration of the vehicle. On the other hand, the information of road surface obtained from the front of vehicle can be detected on line, when the wheeled vehicle runs over the irregular terrain. By means of effctive utilization of front information, it is naturally expected that the better vibration control performance can be obtained.
This paper proposes a vibration model for a four wheeled vehicle with actuator, which presents the displacement in vertical, rolling and pitching directions of the vehicle caused by irregular terrain. Further, vibration control of the four wheeled vehicle running over irregular terrain is considered from the general control point of view. A new vibration control method based on optimal preview servo system is suggested. By use of this method, the vibration in every direction is controlled, and vehicle position can be arbitrarily controlled at the same time. Efficiency of preview action toward the vibration control of the vehicle is also confirmed by means of computer simulation.

抄録全体を表示

As more and more human-related variables such as preference, subjective decision making, sense of taste or smell, and so on, are involved in the discussion of measurement and evaluation, the theoretical exploration into the foundation for measurement from a broad perspective is becoming important.
One remarkable character of human-related variables can be seen in their binary relationhship. Almost always intransitivity appears. For example, from among a set of objects {A, B, C, …}, take a pair of them and ask the evaluator: “Which do you prefer?” (the pairwise comparison) The evaluator may prefer A to B and B to C among many such comparisons. There is often the case, however, when A and C are paired later casually, the evaluator chooses C rather than A! In such a case, the data are represented as an asymmetric directed graph known as “tournament” which has cycles
Here, we present an analysis of the strongly connected tournament, sometimes too complicated and difficult to analyse.
By the operation of “addition of point”, we clarify the structural relationship among strongly connected subtournaments, and show the fundamental properties on strongly connected tournament, which yieled some useful information on transitive subsets in the intransitive structure as a whole.

抄録全体を表示

This paper deals with the estimation problem of a random field with discontinuities. The method is based on the Bayesian approach. When the Bayesian method is applied to the estimation problem of a random field, a prior information about the random field plays a crucial role. It is revealed that in the case of the random field with discontinuities, as the prior information, the Cauchy probability distribution for the spatial variation of the random field is superior to the Gaussian probability distribution which is widely used. Then, it is shown that the estimation algorithm obtained is realized in a parallel manner using a neural network model.

抄録全体を表示

When we apply a hierarchical neural network based on the back-propagation algorithm to a particular problem, we must determine beforehand the suitable size of network for the problem. But it is a very difficult problem. Too small a network will not learn at all, while too large a network will be inefficient and worsen its generalization ability due to overfitting.
In order to solve this problem, in this paper we propose a compact structuring method based on learning with a large size network and then compacting gradually the suitable size network by combining extra hidden units. The result is a small and efficient network that performs better than the original. Also we demonstrate the effectiveness of this method by appling it to two problems, i. e., to identify a logic function and to recognize handwriting characters.

抄録全体を表示

The robust-fault-tolerance is a property that a computational result becomes nearly equal to the correct one at the occurrence of a fault on the input and output of any gates. In this paper, a robust-fault-tolerant parallel adder and a multiplier are proposed for highly safe digital system. Special numerical representation called distributed coding is introduced to satisfy robust-fault-tolerance, in which a numerical value is expressed by the number of 1's. The intermediate results of the arithmetic circuits is almost equal to the total number of 1's in the input data. This property makes linearly additive effect with small weight to the final result in the occurrence of both temporary and permanent faults.

抄録全体を表示