現在IRISは宇宙科学研究所が中心となり,名古屋大学,東京大学,国立天文台,通信総合研究所やその他の研究機関の共同ですすめられている.われわれスペース赤外線観測による天文学・天体物理学に携わってきたものにとっては,初めて世界の第一線に肩を並べるという意味で,ずっと待ち望んでいる計画である.わが国初めての本格的な宇宙赤外線望遠鏡衛星を実現するために,多くの天文学,工学研究者のサポートを期待する.なお宇宙研のWebサイト,http://www.isas.ac.jp/info/future-j3.htmlも参照されたい.

抄録全体を表示

This paper is an investigation into autonomous distributed CAD systems based on the self-organization function of life, which in the real world is a proficient autonomous distributed mechanism. A cell which is a basis unit of life realizes self-organization by repeating simple "stimulus-reaction" cycles (so called cyclic-loop) to emerge upper level function. As design is defined as a building up the each function of a part to satisfy design specification, self-organization function is useful to implement CAD systems. Although autonomous distributed systems have been an object of study, little attention has been given to constructing CAD systems using a cyclic-loop. This "cyclic-loop" process has been formulized and mapped to artificial systems. An artificial autonomous distributed system with a self-organizational function using a modelon system is proposed; it uses a simple linear feedback mechanism and a random search method. This approach is adapted for a "gear layout" design and experimental results are given.

抄録全体を表示

本論文では,力学的なシミュレーション技術を利用した,図形の配置設計の支援手法について議論した.まず図形配置に関する制約を,領域的な制約と寸法制約の2種類に分類し,力学的なシミュレーション技術を用いて,領域的な制約を対話的に充足させる手法を示した.次に寸法制約の多くが,図形問の距離や姿勢を拘束する平面リンク機構と見なせることを説明し,同じシミュレーション技術により,寸法制約の図形の位置姿勢に及ぼす影響が,適切に評価できることを示した.提案した手法に基づいて配置設計支援システムを試作し,手法の有効性を検討した.
作成したシステムは実験的なものであり,実用化のためには以下のような課題が残されている.
(1) 現状では図形を配置する操作として,マウスによるドラッグ操作だけを用意している.Haradaらの提案している離散的な移動など,より多彩な配置操作を実現する必要がある.
(2) 本論文では,領域的な制約に対応する障害物の導出法については議論していない.領域的な制約を整理し,汎用性の高い導出法を確立する必要がある.
(3) 寸法制約は,仮想的なリンク機構を定義したファイルを読み込ませることで入力している,これについても,自然な入力用インタフェースを用意する必要がある.
(4)本システムで最も時間を要する処理は,図形間の接触点の検出である.連続的に移動する図形間のすべての接触点を, ほぼ一定時間で検出できるなどの手法の導入を検討している.
実用性の高いシステムの完成後に,より詳細な評価実験を行いたいと考えている.

抄録全体を表示

In conventional modeling, a design object is mostly described as a static and unique structure composed of entities. Such a static model only gives a snapshot of the design object on the specific aspect. For advanced product design, products should be modeled as both time-changeable and multiphase objects. To deal with machine behaviors, the changes of attribute values and relations of components should be described explicitly according as its history. Industrial products are now considered as multidisciplinary objects. Entities such as parts and units are inadequate to represent design objects with behaviors and multiphase in some integrated way. In order to extend the entitybased modeling framework so as to include behaviors and multiphase, the authors proposed the statebased modeling framework. The modeling primitive in the framework is called 'state'. Each state is described as a combination of the spatial region occupied, the time interval in which it exists, and a list of attribute values for the specified spatial region and time interval. Entities such as parts, units, and products are represented as the aggregation of states which are distributed in the space-time extent.

抄録全体を表示

FDL (Formal Engineering Design Language) has been developed to provide integrated support environment for basic design. It is an object oriented language which uses constraints for its expressions and information exchanges between objects. The date retrieval functionality of relational database, connected to constraint networks, is provided. FDL could be a reasonable parametric design support tool for routine design, however it was not enough for more innovative design support as its model structures could not be modified while it is under operation. To overcome this problem, several meta operations including a structure merging operation, 'was_ a' and 'was' links to manage relations between objects, and configuration operators for automatic model modification were introduced.

抄録全体を表示

Pistons made of copper are used for radiate heart from semiconductor bare chips mounted on a ceramic module board used for main frame computers. This paper proposes a method to inspect locations of pistons using color image processing. After a color image is transformed to a binary image through color space conversion, only pistons are extracted using the copper color. The logical data for color conversion and color extraction are set in a Look-Up-Table for real-time color processing. As the experimental result, reliable inspection was realized which had not been automated by gray-scaled or binary image processing.

抄録全体を表示

For reasons of the various defect and the complicated shape, defective inspection of casting is depened on human eye. A judgement gets unstable by a tiredness of a person and change of an environment. and becomes subjective. So, development of automatic inspection by image processing was tried. Firstly, in order to reduce the noise which occurred by the surface roughness, a method lighted up using the several fluorescent lamp which did ring shape uniformly was used. Secondly, in order to distinguish between a defect and the noise which occurred by the surface roughness, band pass filter by image processing was used. This filter operation improves SN ratio, between superficial noise and defect signal, 1.67 to 5.35. The band pass filter is composed of averaging operators and differential operators. Thirdly, to distinguish defects and pseudo defects. such as characters, steps. edges. dimples and so on, statistical compared method was applied. This is a method compared with the result that was calculated from several good products statistically. using these methods, good products and defective products were perfectly distinguished.

抄録全体を表示

This paper presents a new self-calibration method for displacement or angle sensors which are installed in the instruments. The principle of the method is based on the fact that the linearity error of the target sensor is very small. On every sampling point the derivative of the calibration curve function is measured simultaneously with output of the sensor. Input value of the sensor at every sampling point is approximately estimated first by using a straight calibration line which is given externally. The derivative is integrated with referring the approximate input to get the approximate calibration curve. Then the approximation of input value of every sampling point is improved with the obtained calibration curve. Next the integral of derivative is improved with the newly estimated input values. Repeating these improving process the calibration curve converges to the correct one. Results of simulation and experiment have shown that the calibration error of the method is less than twice of the resolution of the calibrated sensor.

抄録全体を表示

Physicians and nurses must monitor and treat many patients continuously and intensively in an Intensive Care Unit (ICU). To enable these health care professionals to devote more time to critical patients, computer systems must support their treatment of less critical patients. The regulation of water balance between total fluid infusion and urinary volume is an important factor to stabilize a patient's condition after open heart surgery and subarachnoid hemorrhage surgery. This paper describes a water balance computer controlled system utilizing a fuzzy reasoning approach in maintaining the patient's water balance. The system's computer calculates fluid and diuretic infusion rates in balance with urinary volume based on the physician's prescribed input. This paper also describes the system's safety features. One of the safety features is a setpoint changing function for fluid control. Another safety feature is the diuretic injection cessation function when the central venous pressure falls out of the physician's prescribed input range. The system has been used in 23 cases demonstrating that it can regulate and maintain water balance in patients according to a physician's prescribed input.

抄録全体を表示

Measurement error of the CBP, capacitance ball probes, method caused by the floating capacity has been discussed by means of the IEM, integral element method. The CBP method was developed by the authors to precisely measure the circular motion errors of a high precision CNC grinding machine. In this method two precision steel balis, one is the sensing probe connected to the capacitive meter and the other is grounded, are utilized. In the measurement of the circular motion, change in clearance between two ball probes is sensitively detected. The detected data, however, tend to be affected by the floating capacity which is generated between sensing probe and adjacent bodies. The measurement error caused by the floating capacity is almost proportional to the setting clearance between two balls and it sharply increases with the diameter of the sensing probe. When the sensing probe is sandwiched between auxiliary electrodes, distortion of the lines of electric force between two probes is suppressed and the measurement error is successfully diminished.

抄録全体を表示

本論文では,マニピュレータの冗長自由度を利用したコンプライアンス制御の一手法を提案した.作業中にロボットの途中リンクと環境とが接触するような状況に対して,マニピュレータの姿勢に十分な柔軟特性を持たせることで回避できることが数値実験,実機実験により示された.またマニピュレータの途中の点の柔軟度が明確に定義されたことにより,環境に応じたより良い姿勢の評価指標として利用できる.今後,複数点の制御点設定時における階層的な剛性パラメータ設定アルゴリズムの考察,つまりマニピュレータの姿勢の任意点に任意の柔軟特性を指定するアルゴリズムの導出が課題である.

抄録全体を表示

触針に作用する測定力を制御する方式による新しい触針式表面粗さ測定装置の実現を目的として,本研究では粗微動連動制御式表面粗さ測定装置の開発を試み,片持ちはりを用いた触針式プローブの構成と基本特性,測定系の周波数応答などの特性,ならびに種々の表面粗さ測定結果について述べた.
主な結果を以下にまとめる.
(1)粗微動連動制御に用いる触針式プローブを考案し,その有効性を確認した.
(2)試作した粗微動連動制御式表面粗さ計により種々の加工面の表面形状測定を行い,従来の触針式粗さ計によった測定結果と比較することにより本測定装置の有効性を示した.
一方,触針式プローブの出力に重畳する雑音成分,ならびに触針の接触力の低減を図ること,制御系の周波数応答特性を改善して,測定速度を高めることなどが今後の研究課題として残された.

抄録全体を表示

Small particles (approximately larger than 1μm) adhered on to TFT-LCD (Thin Film Transistor-Liquid Crystal Display) plates in a production process cause the defects in circuit patterns. The particles have been detected strictly in a production process by employing automatic inspection systems. In this development, an automatic particle detection system with a high sensitivity and a high speed has been developed. The high sensitive detection has been realized by developing the technology to illuminate a plate with a laser diode obliquely and to detect the forward light scattered by particles obliquely. The high speed detection has been realized by developing a parallel detection system. In this system, the width of detection unit has been reduced by combining an optical path for particle detection with an optical path for automatic focusing. This technology enables to inspect a plate with the size of 360 × 460mm in 24 seconds when the plate is inspected with 4 detection units.

抄録全体を表示

Piezoelectric actuators with high resolution, high stiffness and fast frequency response are widely assembled in micropositioning applications. The tracking accuracy of piezoelectric actuators under open-loop control is seriously limited due to inherent hysteresis nonlinearity in their response to an applied voltage. In a closed-loop control system without sufficient phase margin, the hysteresis will cause instability. This paper presents a computer-based tracking control approach based on polynomial approximation to reduce the hysteresis effect. Because of hysteresis, the expansion of a piezoelectric actuator is dependent not only upon the current value of input voltage but also upon its input history. In this paper, a new parameter called turning voltage is introduced. It was found experimentally that the shapes of hysteresis curve are decided by the value of turning voltage and by whether the piezoelectric actuator is expanding or contracting. The presented approach reduced an inherent hysteresis of 10% to 3% and provided an accuracy of 3% within the working range of 0-4μm under open-loop control. A nonlinear feedback controller based on predicted hysteresis curve compensated the difference between expanding directions in deformation velocity.

抄録全体を表示

The purpose of this study is to design an optimum profile of the flexure hinge with high axial stiffness and high bending flexibility. The conventional circular notched and rectangular notched profiles are not optimum because of low flexibility and high stress concentration. In this paper, the optimum profile with uniform stress distribution was generated using the finite element method under constraint conditions which unify the axial stiffness and minimum thickness of the notched profiles. Moreover, the flexibility of the optimum profile was compared with the flexibility of ellipse profiles, round rectangular profiles and profiles described by higher degree functions. As a result, the optimum profile had flatter stress distribution and the highest flexibility under a maximum stress constraint condition. The optimum hinge has 19% higher bending compliance than the conventional circular notched hinge, and produces 67% higher deflection angle than the rectangular notched hinge.

抄録全体を表示

Single crystalline silicon is one of the most important materials in semiconductor industry because of its superior properties as semiconductor. However, single crystalline silicon is difficult to machine precisely by conventional method because of its fragility. More complex, minute and three-dimensional shape will be required in the near future. From these points of view, electrical discharge machining (EDM) might be used for machining of single crystalline silicon because of very small machining force. In this study, the possibility of EDM for single crystalline silicon is experimentally investigated. The experimental results pointed it out that EDM for single crystalline silicon could be carried out when its resistivity was less than 10^-2Ω ·cm. The removal volume of crater of silicon in a single pulse discharge was much larger than that of steel because of its low boiling point and the important role of Joule's heat generation in material removal. Therefore, the removal rate of single crystalline silicon was much higher than that of SK3. Furthermore, the electrode wear in EDM of single crystalline silicon was lower than that of SK3 and no electrode wear EDM was possible under the appropriate condition with kerosine type fluid.

抄録全体を表示

This study analyzes the time dependent temperature distribution of the wire electrode along the wire axis by the finite-difference method. In the calculation, the heat transfer coefficient between the wire electrode and the working fluid and the energy distribution ratio are exactly determined based on the experimental results. Calculated results show that the mean wire temperature increases with increasing the peak current and discharge frequency, and that the mean wire temperature is higher when the flow rate of the working fluid is lower and the thickness of the workpiece is thinner. Although the mean wire temperature is far below the softening temperature at which the mechanical strength of the wire material falls drastically, the local temperature at the point where a pulse discharge just occurred is highly elevated to the softening temperature of the wire material. As a result, it is considered that the wire rupture can occur even without a discharge concentration when the mean wire temperature is elevated above a certain limit. On the other hand, it was also found that the wire rupture can occur at the point where pulse discharges are successively localized so that the temperature there almost reaches the melting temperature of wire material.

抄録全体を表示

Frictional characteristics of V-ribbed belt of aramid fiber reinforced rubber are significantly affected by form grinding conditions because of much lower coefficient of friction of aramid fiber and higher one of matrix rubber. In this paper, an in-process measuring system of coefficient of friction of V-ribbed belt in form grinding process is developed based on quantitative analysis of coefficient of friction in the previous report. Furthermore, a form grinding technique of V-ribbed belt for controlled coefficient of friction is proposed by investigating coefficient of friction in spark-out grinding process of V-ribbed belt. Main conclusions obtained in this paper are as follows; (1) The in-process measuring system of coefficient of friction of V-ribbed belt which can exactly predict its wide variation in form grinding process is developed. (2) In spark-out grinding process of V-ribbed belt, generation of V-groove is finished in a while from starting of spark-out grinding, and projecting length of aramid fiber continues to decrease gradually until the end of spark-out grinding process. (3) V-ribbed belt which has the required coefficient of friction can be finished by stopping grinding of V-ribbed belt when the measuring system points at the required coefficient of friction in spark-out grinding process after finishing to generate V-groove.

抄録全体を表示

The kaleidoscope is a device to convert the power density distribution of the laser beam. Rectangular pipe kaleidoscopes are applied to laser hardening experiments in the present research in order to show their effectiveness in the laser processing. The shapes and the dimensions of the kaleidoscopes are determined based on the theoretical analysis of the power density distribution of the laser beam converted by the kaleidoscopes. Laser hardening experiments are carried out for graphite-coated steel plates by applying both the kaleidoscopes and the conventional lens system. The heat affected zones of the processed materials are observed and compared from the viewpoints of the depths, the widths and the shapes. The following remarks are concluded from the experimental results. The heat affected zones of the workpieces generated by the laser beam converted by the kaleidoscopes are flatter and wider than those obtained by applying the conventional lens system. The surface deformation of the heat affected zone is relatively smaller in the case of the rectangular pipe kaleidoscopes than the conventional lens system.

抄録全体を表示