A variety of destructive and non-destructive inspection and estimation systems have been developed and used in the steel industry to support the quality assuarance of steel products. In these inspction and estimation, the important problems have been experienced using a microscope to inspect for estimation of non-metallic inclusions of steel products. The non-metallic inclusions inspection is kept severe criterion in wire rods which are making steel cords for automoble tire. Usually the criterion of the wire rods for the steel cords are applied Michelin Method. The harmful inclusions in wire rods for the steel cords are inspected and the size of inclusions are meausured and classified. Actually several automatic systems have been developed to inspect or to estimate of various inclusion of steel products. However, these systems have not been applied to satisify the criterion of the Michelin Method. In order to make the result of visual inspection corresponding to the result of an automatic inspection system, the resolution capacity of image processor system should approach that of the human visual resolution. The automatic microscopic inspection system has been developed using high resolution image processor with new descriminating method among various kinds of non-metallic inclusions. The accuracy of the system is kept within 0.5μm in the defined value. And standard deviation of the system is kept within 0.5μm. The processing time of this system is 3.4 seconds per one frame or 15 minutes for one specimen which have 265 view fields.

抄録全体を表示

This paper describes a new remote sensor for noncontact surface hardness evaluation. Pulsated air jet is generated by acoustic device which drives small area of object surface. The induced vibration is measured with laser interferometry, and the surface hardness is evaluated from it. This process is done without any mechanical contact. The new sensor has some merits over the ones which we have previously reported: firstly, it is less dependent upon surface geometry, secondly, it is not necessary to set the sensor so near the surface, and last, wider rangeability of hardness is realized. The sensor can be applied to robot for soft griping of delicate works, and also to estimation of ripeness of foods such as fluits, or of dryness of painted wet surface. Both sensor and object can get rid of any damage and problems due to mechanical contact in tactile sensing.

抄録全体を表示

We'v already proposed Serial Multi-Point Method (SMPM) for surface profile measurement¹⁾. This method is very powerful because the influences of the sensors head's fluctuations can be cancelled completely and the accuracy also can be improved with only increasing the number of sensors.
But, recursive solution, solution which gives the relationship between two measured results of surface profile before and after the one more step advance of the sensors head, was not established yet.
So in this report, we'v tried to establish recursive solution. And then, using achieved result, we made it clear how the measured result of surface profile will be extended as the sensors head's traverse motion proceeds and the number of measuring points increases.
Consequently, the transparency of theoretical aspect of SMPM improved in great degree.

抄録全体を表示

In the previous paper, the authors proposed the improved double measuring method for calibration in dynamics of linear sensors, and also presented the novel method to identify and compensate the differences in dynamics between two sensors with the aid of the improved double measuring method. From the practical point of view, however, the discussion about the novel method seems to be insufficient.
This paper is devoted to further discussions about the novel method.
Firstly, we extend the measuring method to the case where the sensors are used for simultaneous sensing of the different kinds of physical quantities. Secondly, we consider how to compensate the differences in dynamics between the sensors. Finally, with the experimental results of the dynamic mass-measurement, we demonstrate the validity and applicability of the novel method.

抄録全体を表示

Unknown curved surface can be estimated by a linear combination of a set of adequate basis functions. This paper determines the optimal basis functions for the estimation of the unknown surface of a powder-body for the case where a priori information is available.
The determination is considered from two aspects. The first is the difference depending on the number of the observations available. The second is that depending on whether we consider a pure estimation of the curved surface or an accurate measurement of the unknown volume of the powder-body.

抄録全体を表示

Oxygen saturation can be measured by absorption spectrophotometric method based upon Lambert-Beer law. Mainly two types of oximeters had been developed and applied to the clinical use. One is used to measure oxygen saturation of hemolyzed blood put into a cuvette, and the other to measure arterial oxygen saturation in a tissue like a fingertip. However, since optical characteristics of the micro vessels which exchange oxygen and carbon dioxide is different from those of hemolyzed blood and the tissue, there is no methodology to measure the oxygen saturation in micro vessels. Therefore the knowledge about the changes of the oxygen saturation in each micro vessel is barely reported.
In this study, we employed a special cuvette as an in-vitro model of micro vessel. By comparing cases of illuminated object with parallel light and complete diffuse light, we investigated the effect of light scattering by blood cells and influence of physiological parameters of whole blood (hematocrit, thickness) on the measurements. As results, eliminating the influence of light scattering by the blood cells was considered to be important to precise measurements of the oxygen saturation. We found that the proposed method is not affected by variation of those parameters and showed the ability to measure the oxygen saturation in microcirculation.

抄録全体を表示

In this paper, we present a method of estimating a domain on the complex plane where only one eigenvalue of an interval parameter matrix exists. A sufficient condition for the existence of such an eigenvalue is given based on the Gershgorin's theorem. This condition is represented as an equivalent condition that the convex combination of two matrices becomes an M-matrix. A method for checking the condition is shown by utilizing the property of M-matrices.

抄録全体を表示

This paper considers an H_∞ optimal control problem for SISO systems. The problem treated here is that of obtaining an optimal controller under the constraint that the closed loop poles locate in a specified domain of stability in the complex left half plane. A typical domain of stability is S_δ={f: poles of f(s) in Re s≤-δ}. It is shown that the infimum of the performance index remains unchanged though such a constraint is imposed. It is also shown that the optimal controller exists if δ is small enough. For large δ, however, finite dimensional optimal controllers do not exist. Suboptimal controllers achieving the performance arbitrarily close to the optimal value are constructed by Runge Theorem, togather with providing the explicit bound of the approximation error. The δ introduced here is the distance from the dominant poles to the imaginary axis and is thought of as the degree of stability of the feedback system. It is shown that the large δ, however, gives rise to the decrease of the margin of robust stability for unstructured uncertainty.

抄録全体を表示

This paper is concerned with deadbeat control of continuous-time systems. The objective is to construct a continuous deadbeat tracking system, in which the tracking error settles down to zero in finite time or deadbeatly. It is shown that such a tracking system can be obtained by using controllers which include some delay elements.

抄録全体を表示

This paper presents a design method for the direct model reference adaptive control system which is able to behave robustly in the presence of bounded disturbances. In the method, the controller parameters are adjusted by an adaptive law constructed with normalized signal and dead zone, and the output error between the plant and the reference model is fedback to the control input through a fixed compensator of PI-type to reduce the residual tracking error resulting from the use of dead zone. Stability of the overall system is proved by applying the idea of frozen system, based on the fact that the controller parameters vary slowly with the lapse of time. Simulations are carried out for a simple example system and it is then shown that the proposed method is useful in the sense that the adaptive law guarantees the stable behavior and the fixed compensator improves the control performance not only in the steady state but also in the transient state.

抄録全体を表示

Although the GMDH (Group Method of Data Handling) can be applied easily to identification of nonlinear systems, the interpretation of the layered models obtained is not so easy because of insufficient mathematical background for it. Therefore, the application of the GMDH have been restricted to the field where the model is treated as a black-box.
The present paper proposes a new nonlinear impulse response function with a single time-lag in order to evaluate an overall nonlinear characteristics of the layered model. It can be calculated easily by exciting the layered model with a certain input impulse, and the frequency response function and the power spectrum can also be estimated from it.
As an example, an autoregressive-type nonlinear layered model based on the GMDH was fitted to the vibration data which were generated from a stochastic Duffing equation with a double-well potential. Through the analysis using this new nonlinear impulse response function, it is clearly shown that the obtained layered model contained a mechanism of random jumping of amplitudes between two potential levels.

抄録全体を表示

The CMAC has been paid attention as a neurological model in the hope of achieving human-like performance in a nonlinear controller design. The CMAC can approximate a wide variety of nonlinear functions by training. The training speed is very fast since the computation is local. Therefore CMAC is well applicable to a large scale system. On the other hand, the disadvantage is that CMAC requires an extremely large memory capacity since it is based on a table look-up method. Furthermore, a trade off between a memory size and an approximation accuracy is required in the choice of the quantization interval.
In this paper, we design a CMAC that has a hierarchal memory structure. The memory structure enables a dynamical segmentation of memory cells. In the region where the output of the nonlinear function to be learned changes rapidly, the quantization interval automatically gets smaller so that the approximation accuracy is improved. This technique ensures an efficient memory usage and alleviates a difficulty of the choice of quantization interval.
Taking a learning control of two degrees robot manipulator as an example, computer simulations are performed to show the validity of the proposed CMAC. The simulation results show that the CMAC requires 1/3 less memory than the previous one without degrading the control performance.

抄録全体を表示

This paper describes the fuzzy drive control of an intelligent robot. The fuzzy control is applied to the controls of the steering (direction) and the driving speed of the robot.
This fuzzy control system architecture is a hierarchy that is constructed by a visual sensor subsystem, a motor drive unit, and a fuzzy drive expert system which manages those subsystems. The fuzzy drive expert system is consist of the environment recognition rules, the steering control rules, the speed control rules, and the predictive control rules. One of the feature of this system is that the possible driving area is obtained from the driving results simulated by the candidate pairs of directions and speeds prior to the drive control. It was called the predictive control. The directions and the speeds for the drive control are decided by unifying the candicate pairs for directions and speeds in the possible driving area and those from the steering-speed control.
As an application to the fuzzy drive control, the driving experiments of an autonomous mobile robot on the straight road and the one with corner were implemented. From the experimental results, the usefulness of this fuzzy drive control system was recognized.

抄録全体を表示

A guided flight vehicle using the principle of proportional navigation has a target tracker to measure the line-of-sight angle rate. The tracker dynamics have the uncertainties caused by the gain variation in tracking loop, the imperfect tracker stabilization and the tracker acceleration sensitivity.
In the guidance system including these uncertainties, the target tracker measurement noise such as white glint noise and white fading noise is apt to aggravate the system performance.
Applying Adjoint Method to this guidance system, we obtain a new performance index, which can be used to evaluate the uncertainties from the standpoint of system performance. In addition, the relation between the index and the system performance is presented.

抄録全体を表示

A reasonable formulation of Petri net construction problem and the solution of the problem are important to the rational development of discrete event systems. This paper deals with the following condition/event-net (C/E-net) construction problem: given a partial language over a set of events as specification, construct a C/E-net which generates the same partial language as the specification using the only events described in the specification. It first shows a necessary and sufficient condition for the problem to be solvable. It next discusses the properties of solutions and then presents a method of finding solutions. The method constructs a C/E-net with less redundant conditions.

抄録全体を表示

In the previous paper, we proposed an effective spectrum analysis method for transcription, in which adaptive technique was used. Frequencies analyzed by the FFT method are integral multiples of one particular frequency, but in this proposed method frequencies can be selected freely, so analysis along the musical scale without frequency correction is possible.
In this method, LMS adaptive algorithm is used. The step size parameter is a convergence factor that controls stability and adaptation rate. In this paper, we derived a method to decide the step size parameter from the condition for convergence of the coefficient and the learning curve, and from the relation between the time constants and the duration time of sound. Simulated and experimented results are illustrated.

抄録全体を表示

An approach is presented for increasing the speed of parameter convergence in adaptive control systems. In this approach, the adaptive law is modified by using a prior knowledge of plant.

抄録全体を表示