巻号一覧
31 巻, 8 号
選択された号の論文の39件中1～39を表示しています
- |<
- <
- 1
- >
- >|
-
北山 正文1995 年 31 巻 8 号 p. 951-956
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーThe particle size distribution, and particularly the particle shape, of pulverized coal dominate the coal combustive condition.
No accurate measurement of a particle volume has been developed yet and volumetric shape factor of a fine particle has not been obtained.
This paper proposes a novel measurement of volume of a fine particle using three dimensional image processing. And the accuracy of measurement are discussed on the basis of simulations using polyhedrons.
In this result, the measurement errors were less than 16%, when we chose the appropriate lens. So the method is found to be applicable to analysis of pulverized coals.
And we show that equivalent volume diameter estimated by the technique is identical with the mesh size of seive.抄録全体を表示PDF形式でダウンロード (1704K) -
北山 正文, 河野 昌史1995 年 31 巻 8 号 p. 957-962
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーAt present, two phase flow of air-pulverized particles (several tens micrometers-several thousands micrometers) can be seen in the equipments used by the industry for producing some kinds of powder, medicine, and so on. The on-line measurement of the particle size distribution is necessary to control the equipment used in the manufacturing processes of particles. In this paper, using the mathematical model and experiments, it will be confirmed that the uniform phenomenon of falling velocities of particles in dilute two phase flow exists. The fundamental study will privide for a fast and simple measuring method of particle size distribution in two phase flow of air-pulverized particles. And, it will be shown that this fundamental study is useful to measure the particle size distribution using the value of the staying time of the flowing particle in a semiconductor laser beam.抄録全体を表示PDF形式でダウンロード (2427K) -
金 合烋, 小野寺 聡, 浅野 春生, 須田 昌克1995 年 31 巻 8 号 p. 963-972
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーA novel approach is proposed to digitizing series integrators which are often used in modeling dynamical system and in realizing controller and filiter. The so-called causal signals are treated, and sampling techniques in the time domain are availed. Using the z-transform, the closed expression is presented which states equivalence between analog and discrete systems for series integrators. This follows that every analog integrator in a series connection has an equivalent discrete description and vice versa. The procedure is given for the implementation of the every digital integrator deduced from the description. And, an approximation method is proposed to reducing the number of operational elements employed in the implementation of higher order cases. The method is done by using a delayed integrator placed at the first stage and bilinear integrators arranged at the following stages in the digital series connection, respectively.
The approach is applied to digitizing the Lagrangean B-filter being formed of feedback system. The impulse response of the mth-order B-filter holds a non-negative even function symmetry, and all its derivatives of order up to (m-1) are continuous. Hence, the B-filters have linear phase properties in nature. Therefore, a novel type of the recursive digital filter added interpolating operations derived from the B-filter takes perfectly linear phase.
Computer simulations have verified the applicability of these techniques. Those are illustrated by some numerical examples.抄録全体を表示PDF形式でダウンロード (2551K) -
源野 広和, 松本 和夫, 福島 清司1995 年 31 巻 8 号 p. 973-981
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーThe facial skin temperature can be measured without contact using an infrared camera regardless of room temperature conditions. Moreover, because it varies with autonomic nervous activity that accompanies arousal of the senses, it serves as an effective physiological quantity for evaluating human senses.
This report examines the thermal sensation and evaluates it using the facial skin temperature without influence by stress or the change in seasons. In the experiments that were carried out under various types of stress with various types of ambient temperatures in winter and summer, no change due to seasons was found. Consequently, data for both winter and summer were combined and correlation coefficients between the thermal sensation and facial skin temperatures for various parts were determined. In the experiments, the correlation coefficients decreased in the order of cheek, chin, forehead, nose and ear. Even the lowest value of r=0.752 for the ear is still high. In addition, the thermal sensation is estimated by a multiple regression formula using the stepwise method. The partial regression coefficient for the nose excluding the forehead became negative (multiple correlation coefficient R=0.897). Accordingly, the relation between the thermal sensation and the forehead skin temperature can be predicted by the skin temperatures of the other facial parts, and a rise in the thermal sensation accompanying stress can be described by a drop in the skin temperature of the nose.
Furthermore, it is suggested that this multiple regression formula is applicable even for different amounts of clothing. The compensation is necessary for changes in the activity level and continuous variations in the ambient temperature.抄録全体を表示PDF形式でダウンロード (3110K) -
Analysis of Basic PerformancesIvan GODLER, Akira AKAHANE, Toshiki MARUYAMA, Tadashi YAMASHITA1995 年 31 巻 8 号 p. 982-990
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーAcceleration is frequently used in motion control. It is usually obtained as a time derivative of position, which is measured by an encoder. A time derivative is corrupted by high frequency noise and is not reliable at low rotational speed. Since there is no appropriate angular acceleration sensor available at present, we propose a new type of angular acceleration sensor which has unlimited rotational sensing range and sufficient sensitivity to be used in motion control. The proposed sensor is composed of two discs. One of them is a spring-inertia acceleration transducer and the other is a disc similar to the discs of pulse encoders. A pair of optical pick-ups is used to measure a relative angular displacement between the two discs. In this way non-contact and speed independent angular acceleration sensing is realized. Some basic performances of the sensor are discussed in this paper: it is shown that linearity of the sensor can be improved by slits with a shape of Archimedes' spiral; eccentricities of the shaft and the discs are analyzed and a sufficient but not ideal compensation is provided by two optical pick-ups that are placed diametrically on the opposite sides of the discs. Experimental results show that the sensor's output excellently matches with the calculated acceleration.抄録全体を表示PDF形式でダウンロード (1953K) -
石橋 雅裕, 高本 正樹, 中尾 雄一, 横溝 利男1995 年 31 巻 8 号 p. 991-998
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーVenturi and quadrant nozzles of a total number of 27 were made by using super-accurate lathes, positions of the cutting heads of which were controlled with a resolution of 0.1μm. Their discharge coefficients were measured in the critical condition by using a constant volume tank system at the Reynolds number of 0.8-2.4×105. All the measured discharge coefficients on the Venturi nozzles distributed along a fitted curve as a function of the Reynolds number with a deviation of less than ±0.03%, which showed a feasibility of an accurate standardization on critical Venturi nozzles made by the super-accurate lathes. The measured discharge coefficients on the quadrant nozzles were larger than the fitted curve by 0.07%. The discharge coefficient of a critical Venturi nozzle was theoretically derived into an analytical form from theories of an axis-symmetric inviscid flow field and a laminar boundary layer at the throat of a critical nozzle. The calculated discharge coefficient agreed with the fitted curve with a deviation of less than 0.03% at the Reynolds number larger than the lower extreme of the measuring one.抄録全体を表示PDF形式でダウンロード (1437K) -
Taiichi TESHIMA, Satoshi HONDA, Yutaka TOMITA1995 年 31 巻 8 号 p. 999-1004
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーMagnetic flowmeters, which have been one of the standard instrument to measure liquid flowrate in industry, were applied to tomographic imaging of flow velocity field, based on the fact that there are no exciting magnetic fields that induce the flow signal proportional to volumetric flowrate free from the velocity profile. Two-dimensional case was examined theoretically for the reason of simplicity, and a flowmeter with eight magnetic coils and eight signal pick-up electrodes was developed. Driving alternating current of unit ampere was fed to eight combinations of pairs of the coils and induced electro-motive force were picked up across eight combinations of pairs of the electrodes. Flows in a pipe were assumed to be fully developed and axially symmetric since the straight approach flow has the length of forty-five times of the pipe diameter (45D). Since field interaction model of the meter is ill-posed, some prior knowledge, such as non-negativeness of the flow velocity and smoothness of the profile, was introduced to reconstruct the velocity profile.
In order to examine the possibility of reconstructions for various velocity profiles, to make different velocity profiles is needed. When the pipe is shifted perpendicularly to its axis, the weight function is different from what it was. Therefore the shift of the pipe position was treated instead of the different velocity profiles for simplicity. As a result, when the pipe was shifted, the same reconstruction regardless to displacement was achieved. The flowrate estimated with the present equipment agreed with a commercially available standard magnetic flowmeter within 5% accuracy.抄録全体を表示PDF形式でダウンロード (1828K) -
松岡 祥隆, 山本 芳巳, 田辺 正則, 山田 一二, 嶋田 智, 保川 彰夫, 松坂 英俊1995 年 31 巻 8 号 p. 1005-1014
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーThis paper describes temperature dependence of an output voltage nonlinearity on a silicon diaphragm pressure sensor utilizing the piezoresistance effect for high pressure measurements of a 50MPa span. The diffused gage resistor has temperature dependence of the nonlinearity and it is necessary that the nonlinearity of the pressure sensor fabricated in a pressure transmitter for industrial use is within ±0.2% and an output voltage/supply voltage is above 43mV/V at the same time.
When stresses on the silicon diaphragm are analyzed by a plate theory of fixed periphery of a classical dynamics and a finite element method (FEM), the stresses calculated by the latter shift about 125MPa for compressive side from those by the former. A value of the gage resistor diffused on the silicon diaphragm is calculated by an equation which is expressed quantitatively as a function of the tri-axial stresses by FEM, the high degrees piezoresistance coefficient and its temperature coefficients obtained by our previous experiments. When the calculated results of the nonlinearity are compared with the experimental results, we found that stresses calculation by FEM is more effective than the one by the plate theory of a fixed periphery in order to forecast the experimental results.
We found from the experimental results that there is the gage resistor configuration within ±0.1% of a nonlinearity at a measuring span pressure of 50MPa under temperature range of -40-120°C, and there is also the gage resistor configuration whose temperature dependence of the nonlinearity within 0.02% can be neglected.抄録全体を表示PDF形式でダウンロード (3420K) -
出口 光一郎, 多田羅 哲夫1995 年 31 巻 8 号 p. 1015-1022
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーThis paper proposes a fast and approximate solution of camera calibration for image synthesis. We present three types of the approximation for perspective imaging, and compare their performances to that of the exact solution by exact perspective model. The proposed techniques reduce the processing time drastically. This is because, while the exact solution must employ biquadratic equations, by using approximations to perspective, a fast solution with quadratic equations was constructed for real time image synthesis. This technique allows some amount of errors by the approximation, but it is shown to be tolerant for practical use.抄録全体を表示PDF形式でダウンロード (1723K) -
永戸部 一孝, 高橋 誠, 加藤 徹, 木村 真弘, 伊福部 達1995 年 31 巻 8 号 p. 1023-1029
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーVirtual reality technology provides an interface between man and computer. To make more real virtual environment, it is essential to investigate the ‘Sensory Motor Coordination.’ In this paper, indicating movements and gazing movements were used as one of the ‘Sensory-Motor Coordination.’ The indicating movements, which were the response of normal subjects to optic and acoustic targets, were measured.
The measuring system was composed by a display unit, an eye and head measuring unit and an indicated point measuring unit. The display unit presents targets in a transverse plane of eyes. 19 targets (LEDs and speakers) were arranged from left to right, with separation of 10 degree. Targets were placed at 50 centimeter distance. Stimulation period of the target was 200ms. This stimulation is the same period as the latency of the saccadic eye movements. The eye and head measuring unit was composed by two instruments. One is EOG (electro-oculography). The other is ‘3SPACE FASTRAK’ (POLHEMUS Co.). Head movements were measured by ‘3SPACE FASTRAK’ (POLHEMUS Co.) that uses electromagnetic search-coil technique. Indicating point measuring unit that used a touch panel measured azimuth and elevation of indicated points. The sampling frequency of this total experiment system is 40Hz. Subjects continuously fixated an LED presented straight ahead (0 deg); proper fixation was monitored by an operator. Targets for pointing were presented on the display in a random sequence at eight positions. In a later data inspection, responses were discarded when the eyes had drifted or saccaded during the arm movement.
From the responses of each movement, A shift and S. D. of indicated points increased according to a target presenting angle. This outcome revealed that the response in the peripheral visual field is underestimated on the ‘Sensory-Motor Coordination.’抄録全体を表示PDF形式でダウンロード (1896K) -
舒 志斌, 渡部 慶二, 山田 功1995 年 31 巻 8 号 p. 1030-1036
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーIn this paper, a unification of model and compensator of internal model control systems is presented. It is shown that (1) if the plant does not contain any unstable zeros, the design method of unification of model and compensator on the basis of state space design method of filtered inverse systems is proposed. (2) In the presence of unstable zeros, the solution to the problem is presented by using inner-outer factorization method. (3) By pole assignment, the simple design method of unification of model and compensator of internal model control systems in the SISO system is presented. In addition, simulation result is shown.抄録全体を表示PDF形式でダウンロード (853K) -
Toru ASAI, Shinji HARA1995 年 31 巻 8 号 p. 1037-1046
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーThis paper proposes a necessary and sufficient condition for the quadratic stabilization for a class of linear systems described by descriptor form representations. The system representation contains time-varying but bounded parameters representing the uncertainties. Similarly to the state-space form case, the tabilizing controller can be designed by the H∞ control scheme. We show that the proposed method gives better performances as compared with the state-space method.抄録全体を表示PDF形式でダウンロード (2082K) -
藤田 政之, 丸山 章, 川端 昭弘, 内田 健康1995 年 31 巻 8 号 p. 1047-1053
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーThis paper is concerned with the finite horizon H∞ filtering problem of linear discrete-time, time-varying systems using a direct approach based on the completing-the-square. We derive a necessary and sufficient condition for the existence of the H∞ filters, and give the parametrization of all linear H∞ filters. In particular, the time-updating H∞ filtering algorithm is explicitly shown. Furthermore, we investigate a possibility of the real-time application of the H∞ filters to a robotic visual tracking system. The simulation results show that the derived H∞ filtering algorithm is quite effective for the robotic visual tracking applications.抄録全体を表示PDF形式でダウンロード (1590K) -
柏木 濶, 孫 也平1995 年 31 巻 8 号 p. 1054-1060
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーIn this paper, a new method for obtaining Volterra kernels of a nonlinear system by use of pseudorandom M-sequences and correlation technique is proposed and some applications for nonlinear system identification are shown.
Most of nonlinear systems can be represented by Volterra series. M-sequences are applied to a nonlinear system and the cross-correlation function between the input and the output displays not only the linear impulse response of the linear part of the system, but also crosssections of the high-order Volterra kernel of nonlinear system.
Simulations are carried out for system which contains up to 3rd order Volterra kernels. The results show a good agreement with the theoretical considerations.抄録全体を表示PDF形式でダウンロード (1420K) -
線形行列不等式条件とポリトープ型システムへの適用和田 光代, 池田 雅夫, 太田 有三, Dragoslav D. ŠILJAK1995 年 31 巻 8 号 p. 1061-1069
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーParametric absolute stability is considered for multivariable Lur'e systems consisting of linear parts with uncertain parameters and sectorial bounded nonlinearities. The change of the equilibrium state is dealt with, which is caused by the variation of parameters and reference inputs. Linear matrix inequalities (LMIs) are derived, which constitute a sufficient condition for parametric absolute stability. They are to be satisfied for each value of the parameter vector and a specified class of nonlinearities. One of them ensures the existence of an equilibrium and is used to compute a region where it exists. Others ensure global asymptotic stability of any equilibrium in the region. For Lur'e systems with polytopic linear parts, non-parametric LMIs are presented. They are more conservative than the parametric ones, but can easily be checked via an existing tool for computation.抄録全体を表示PDF形式でダウンロード (1607K) -
堀口 和己, 西村 卓也, 長田 朗1995 年 31 巻 8 号 p. 1070-1079
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーThis paper proposes a model reduction method for MIMO unstable linear discrete-time systems. The proposed mothed provides the reduced-order models which have the following properties. (i) They have the same number of unstable poles as the original system, and (ii) they preserve (interpolate) information of the original system associated with the transfer characteristics and the power spectrum characteristics at appropriate complex frequency points. The model reduction mothod satisfies a reasonable and important request of (i) for reducing unstable systems, and can consider a frequency weight in approximation from (ii). These properties will enable us to make good use of the proposed model reduction method, especially, for controller reduction.抄録全体を表示PDF形式でダウンロード (1432K) -
渡辺 隆男, 安田 恵一郎, 横山 隆一1995 年 31 巻 8 号 p. 1080-1088
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーIn this paper, a reduction problem of H∞ controllers is considered. The problem treated here is that of truncating high order H∞ controllers which are complicated by high order free parameters. This problem sometimes appears when a numerical approach is used for H∞ control problems. The H∞ controller is truncated by approximating a free parameter which is high order solution of the Model Matching Problem using the balanced realization approximation technique. This method preserves internal stability of closed loop systems and the domain in which closed loop poles locate. A sufficient condition for preserving an H∞-norm constraint of the closed loop system in which a high order controller is approximated by low order one is derived. Thus, in this reduction method closed loop properties are taken into account, i. e. the internal stability and the domain in which all closed loop poles locate are preserved.抄録全体を表示PDF形式でダウンロード (1366K) -
Hiroshi ITO, Hiromitsu OHMORI, Akira SANO1995 年 31 巻 8 号 p. 1089-1097
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーIn this paper, we consider performance robustness with two objectives of finite-dimensional, linear time-invariant systems having structured norm-bounded uncertainty, where two performance objectives are given by H∞ norm constraints on transfer functions associated with a single closed-loop system. A methodology for the robust control design is proposed. We prove that a performance robustness problem with multiple objectives is equivalent to a stability robustness problem with extra repeated uncertainty blocks. We show how the equivalent transformation can be exploited to propose a method of controller designs. A numerical example illustrates how the design approach proposed deals with trade-off between two objectives effectively and finds satisfactory controllers achieving two performance objectives robustly.抄録全体を表示PDF形式でダウンロード (1627K) -
陶山 貢市1995 年 31 巻 8 号 p. 1098-1105
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーThe objective of this paper is to establish a new meaning of reliable control with passive redundancy from a viewpoint of cooperation between compensators by introducing a new idea of cooperation in respect of energy.
In this paper, “load” of a compensator is considered to be its output signal against an external command input (e. g., a reference signal). It is evaluated by its total energy (i. e., L2-norm). Then, a new idea: “reliable load-sharing” is introduced as a desirable situation where “loads” of all compensators are always reduced by operating together with other compensators.
It is shown that reliable control and reliable load-sharing can simultaneously be achieved by a configuration of compensators. As a result, a new meaning of reliable control can be established from a viewpoint of cooperation between compensators抄録全体を表示PDF形式でダウンロード (1284K) -
渡辺 桂吾1995 年 31 巻 8 号 p. 1106-1113
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーThe functional reasoning and the simplified reasoning, which are special cases of the so-called min-max-centroidal method, have been applied to construct a fuzzy controller or to model a system. It is not necessary known, however, for control engineers to rationally and easily design the conclusion part or any parameters in the conclusion part. In this paper, a new fuzzy reasoning is proposed so that the conclusion consists of a function of mean-values on each membership function in the antecedent. It is then shown that any parameters except for the mean-values in the conclusion can be rationally and easily designed as a slope of stable switching plane (or line) which is basically independent of the control rule, if the conclusion is formally regarded as a VSS (variable structure system) controller with the meanvalues in the antecedent as states. Thus, the laber required for designing the conclusion part can be reduced drastically by using the proposed method, compared with the traditional methods. The effectiveness of the proposed method is illustrated by simulating a fuzzy controller with two inputs-single output reasoning or a fuzzy controller with three inputs-single output reasoning.抄録全体を表示PDF形式でダウンロード (2595K) -
星野 郁弥, 阿部 禎一, 木村 紘, 木村 英紀1995 年 31 巻 8 号 p. 1114-1121
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーThe new roll eccentricity control system which is an application of the synthesis methodology developed by Kimura based on the design theory of output regulations essentially due to Wonham has been developed and its effectiveness has been demonstrated by the application results to a real plant. The salient features of the new system are as follows.
(1) The roll balancing force regulator is applied effectively instead of the hydraulic roll position regulator.
(2) The controller obtained from the observer-based design method copes well with the actuator and detection time delay without complication of the controller.
(3) The use of the so-called gaugemeter equation for the roll eccentricity estimation leads to the feature that gain tuning with the rolled material variation is no need.抄録全体を表示PDF形式でダウンロード (1258K) -
斉 潤徳, 辻 輝生1995 年 31 巻 8 号 p. 1122-1128
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーThis paper presents an H∞ control system design for a levitation transport equipment by electromagnets attraction with nonlinear characteristics. The control systems consist of a levitation system, a guide system and a run system. As the feature of this system, state variables and outputs evolve in continuoustime and others evolve in discrete-time in the system, and controllers are restricted to be digital controllers through samplers and zero order holds with a fixed sampling period. In the control system design, we consider the influence of both disturbances and uncertainties in the model. The main disturbances stem from the position sensors. The electromagnetic uncertainties are due to the gain change in the current amplifier, the influence of leakage flux and linearization error in the magnetic circuit, and the mechanical uncertainties are due to the changes of the mass and the moment of inertia of the vehicle. The controller design is based on the H∞ type control problem for sampled-data control systems.
The results of several experiments show that the H∞ controller designed for sampled-data control systems gives a higher performance as the sampling period gets longer compared with the H∞ controller designed for continuous-time control systems.抄録全体を表示PDF形式でダウンロード (1280K) -
吉川 恒夫, 原田 研介, 松本 淳, 村上 弘記1995 年 31 巻 8 号 p. 1129-1135
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーIn this paper, hybrid position/force control of flexible manipulators is proposed. To realize the hybrid position/force control, a macro-micro system consisting of a flexible macro manipulator and a rigid micro manipulator attached at the tip of the macro part is used. The macro manipulator can move widely, but cannot realize the fast and precise motion because of its flexibility. On the contrary, the micro manipulator cannot move widely, but can realize the fast and precise motion. In this macro-micro system, the macro manipulator is controlled to roughly realize the desired position and force by a simple PD feedback, and the micro manipulator compensates the position and force errors caused in the macro part. First, kinematic equations of the macro-micro manipulator system are introduced. Second, utilizing the redundancy and the flexibility of the system, trajectory planning is discussed. Third, a quasi-static hybrid position/force control algorithm is proposed. Lastly, to verify the effectiveness of the proposed control algorithm, experimental results are shown.抄録全体を表示PDF形式でダウンロード (1641K) -
原 進, 吉田 和夫1995 年 31 巻 8 号 p. 1136-1143
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーIf influences of such uncertainties as perturbations and disturbances on controllers possess nonstationary characteristics, it is important to develop a nonstationary robust control method. This paper presents a nonstationary robust control method taking account of the nonstationary uncertainty influences by using a time-varying criterion function. In order to improve the robustness of the control, in this method, an integrated design of both H2 control used under the condition of little influences of uncertainties and H∞ control used under the condition of large influences are realized by applying a smooth change of the two kinds of controllers and solving time-varying Riccati equations in stead of time-invariant Riccati equations.
In this paper, as an example in which influences of uncertainties possess nonstationary characteristics, a positioning and vibration control problem of flexible structures is adopted. In this problem, influences of such uncertainties as Coulomb friction and parameter variations are generally severer in the period after positioning than in the period during positioning. In comparison with the former methods proposed by authors, the usefulness of the method is verified by numerical calculations and experiments by its application to a positioning problem of the two-degree-of-freedom flexible structure.抄録全体を表示PDF形式でダウンロード (1740K) -
金子 修, 川谷 亮治1995 年 31 巻 8 号 p. 1144-1151
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーA servo system for any periodical signal can be realized based on the so-called repetitive control system. If a controlled plant has no direct path, the modified repetitive control system which includes a lowpass filter to assure the internal stability, is usually used. As a selection of the lowpass filter is constrained by a gain characteristic of sensitivity function, the sensitivity function must be shaped corresponding to a lowpass filter selected from a standpoint of performance. In this paper, we propose a new synthesis of repetitive control system using the Loop shaping design procedure. This procedure is one of method that can shape the sensitivity function as exepected. Furthermore, a pre-controller designed by this procedure has a structure of state feedback plus full state observer. In a case that plant is of high order, this feature makes a synthesis easy. We consider a periodical disturbance reduction problem in a plate temperature control system and confirm the validity of our method in both simulation and experiment.抄録全体を表示PDF形式でダウンロード (1329K) -
金子 真, 西原 敏晴, 辻 敏夫1995 年 31 巻 8 号 p. 1152-1159
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーUltrasonic Motor (USM) has the capability of maintaining the torque continuously against an environment even under the switch-off. This characteristic is well-known as self-sustaining one and, thus, USM can produce “lock state”. “free state” can be also generated by switching CW and CCW commands periodically with the same time interval. In this paper, we newly propose Two-Parameters-PWM Control, in which OFF period is included between CW and CCW commands. This control contributes to changing the apparent friction between rotor and stator. For example, setting a large OFF period makes the friction increase, and as a result, the rotor becomes difficult to slip against an external torque. As an extreme case, the lock state can be generated.抄録全体を表示PDF形式でダウンロード (2260K) -
笠原 雅人, 松葉 匡彦, 橋本 幸博, 遠藤 龍司, 金原 昭臣, 神村 一幸, 黒須 茂1995 年 31 巻 8 号 p. 1160-1167
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーIn this paper a modeling has been applied to an air-conditioning system by taking the multivariable AR (autoregressive) model into account by using the observed data and the simple 3input/2output model for the implementation of a new practical optimal control has been constructed. The LQG (Linear Quadratic Gaussian) optimal control has been applied to control process variables such as indoor-temprature and indoor-humidity.
The experiments on a commercial sized test plant suggest that the LQG control is extremely good for the air-conditioning system such as the interaction system.抄録全体を表示PDF形式でダウンロード (1671K) -
登内 洋次郎, 坪内 孝司, 有本 卓1995 年 31 巻 8 号 p. 1168-1174
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーThe Kalman filter yields the optimal estimate in the sense of the minimum error variance when all the model parameters are exact. However the model parameters are not always exact in practical cases. The present authors focus on a treatment of Kalman filtering when there are erroneous bias terms in an observation model.
In this case, it is pointed out that the Kalman filter yields neither optimal nor unbiased estimate when it is constructed according to an erroneous observation model. Then, under the existence of errors of the bias terms an estimator is constructed by using a conventional method, where a state vector is augmented with the bias terms. The estimator is proved to be statistically optimal in two different senses. By a simple numerical simulation it is shown to be robust against the errors of the bias terms.抄録全体を表示PDF形式でダウンロード (1696K) -
Toward an Active Recognition of Cursive Handwritten CharactersYoussef IGUIDER, Makoto YASUHARA1995 年 31 巻 8 号 p. 1175-1184
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーWe propose an analysis method of human handwriting movements aiming at an active recognition of cursive handwritten characters. The analysis is performed by synthesizing signals according to a model and comparing the signals synthesized with the signals to be analyzed. To begin with, we will introduce the formal concept of the active recognition. And a model of human handwriting will be presented based on an Auto-Regressive Moving Average (ARMA) process by limiting discussions exclusively to the case of cursive handwriting. The model was first proposed by one of the authors in 19837), and the extensive experimental studies had been made to reveal its utilities. Based on the ARMA model, the mathematical techniques to realize the active recognition will be given by showing that the model reference adaptive control system (MRACS) is able to be applied successfully to extract control signals of the cursive handwriting movements. The control signals are considered to play the most important role in the active recognition of the cursive scripts.抄録全体を表示PDF形式でダウンロード (3431K) -
玉置 久, 森 正勝, 荒木 光彦1995 年 31 巻 8 号 p. 1185-1192
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーIn this paper, a method of applying the Genetic Algorithm (GA) to multi-criteria optimization problems is proposed. First, the reproduction operator in GA is made so that the selection is performed in parallel according to each criterion and, at the same time, the Pareto-optimal solutions are always reproduced if such appear in the population. Then, through computational experiments, it is shown that not only one of the Pareto-optimal solutions of a problem but a set of such solutions is obtained by a single run of the proposed method.抄録全体を表示PDF形式でダウンロード (1466K) -
玉置 久, 西川 〓一1995 年 31 巻 8 号 p. 1193-1201
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーRecently, as the diversification of demand and the progress of production technology result in a wider variety of products, the necessity of scheduling and inventory control is increasing. This paper deals with methods of solution for a job-shop scheduling problem considering the capacities of in-process buffers. To this problem, the authors have proposed three kinds of ways of modeling the problem, i. e., a disjunctive graph (DGR) model, a timed Petri net (TPN) model and a Gantt chart (GTC) model. The present paper proposes several methods of solution based on these models. First, to the DGR model, a conventional algorithm, e. g., a branch-and-bound method, for the ordinary job-shop scheduling problem is applied. Second, to the TPN model, a simulation-based algorithm on TPN, i. e., an algorithm to obtain an optimal schedule among non-delayed schedules by using the branch-and-bound method is proposed. Further, the TPN model is extended to be fit for the problem in repetitive production processes. Third, to the GTC model, both an algorithm to obtain an optimal schedule by using the branch-and-bound method and an algorithm to obtain an approximately optimal schedule by using the list-scheduling method are proposed. Finally, results of several computational experiments are presented, in which the effects of buffers on schedule are investigated numerically.抄録全体を表示PDF形式でダウンロード (1081K) -
佐々木 慶文, 亀山 充隆1995 年 31 巻 8 号 p. 1202-1210
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーIn intelligent robots, three-dimensional (3-D) instrumentation and object recognition are required to move autonomously. Moreover, not only throughput but also the latency between the image acquisition and the output generation of the 3-D instrumentation and object recognition result should be considered because of its frequentt visual feedback. However, enormously large computation power is required to perform 3-D instrumentation and object recognition because they handle two-dimensional (2-D) and 3-D image. Development of a special-purpose VLSI processor for 3-D instrumentation and abject recognition is essential for real-time motion of the intelligent robots.
In this paper, a Model-Based Robot Vision (MBRV) VLSI processor for 3-D instrumentation and object recognition is proposed. The principle of its algorithm is model matching between an input image and 3-D models. Because the algorithm always gives the candidates for the accurate 3-D instrumentation and object recognition result with simple and regular procedures, it is suitable for the implementation of the VLSI processor.
Highly parallel architecture is employed in the VLSI processor to reduce the latency. Pipeline architecture is very useful to reduce the latency because a large amount of data is streamed into the VLSI processor successively. Spatial parallel architecture based on data flow graph (DFG) is also introduced for regular pipeline data flow.
As a result, it is confirmed that the proposed VLSI processor can perform 3-D instrumentation and object recognition 10000 times faster than 28.5 MIPS workstation.抄録全体を表示PDF形式でダウンロード (2312K) -
一軸回転対称物体の場合櫻井 俊和, 大西 昇, 杉江 昇1995 年 31 巻 8 号 p. 1211-1216
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーThis paper presents an automatic generation method of mechanical assembly sequences from geometric information. The obtained sequences represent how to assemble plural parts which have the same axis of symmetry. The drawing used in assembly planning is a cross-sectional view which includes the axis of symmetry, and it is represented in the forms of pixel map. The system consider the disassembly sequences in place of the assembly sequences to reject unfeasible partial assemblies. Thus the system calculates all the possible partial assemblies by iterating an operation of “disassembling one part” starting from the completed object. Since these partial assemblies equal those used in feasible operations represented in an AND/OR graph, the system can generate the AND/OR graph by solving combinatorial problem of the partial assemblies. The assembly operation is feasible, if two partial assemblies used in the operation have no common part and new partial assembly generated from the operation is one of the feasible partial assemblies which have been calculated before. From the AND/OR graph the system searches a tree with the smallest criterion value, which is the best assembly plan. As a result of simulations we have found that the ratio of the feasible partial assemblies to the considerable partial assemblies gets smaller as the number of parts increases. So it is very useful to calculate the feasible partial assembly previously.抄録全体を表示PDF形式でダウンロード (2885K) -
阪口 豊1995 年 31 巻 8 号 p. 1217-1226
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーWhen perceiving an object by touch, human beings integrate various sensory information with observing the object's surface in appropriate manners, such as rubbing, pushing and picking, according to the proceeding of the recognition. This fact suggests that sensory integration and active perception play essential roles in the haptic recognition process.
In light of such characteristics of haptics, the author constructed a haptic recognition system which descriminated feel of touch in a similar manner to human's. The system is equipped with several sensor devices, including a vibration sensor, a friction sensor and a thermal sensor, and can push and rub the object's surface with several values of force and speed. It integrates information from these sensors iteratively with selecting an appropriate sensor and a measurement condition according to the proceeding of the recognition. The algorithms of sensory integration and of active perception are realized by Bayes inference and by an iterative experimental design based on an information criterion, respectively.
The experimental result shows that the system can discriminate subtle difference in feel of touch: It can discriminate 16 kinds of paper and cloth almost perfectly. It is also proved that the system selects appropriate sensors according to the proceeding of the recogntion, that is, the active perception algorithm realizes good recognition accuracy by fewer observations than the random observation algorithm. In addition, it is shown experimentally that the characteristics used in the system well correspond to those human beings utilize in haptic recognition. These results suggest that the constructed system is a faithful model of the human haptic mechanism.抄録全体を表示PDF形式でダウンロード (4027K) -
三宮 信夫, 長野 耕治1995 年 31 巻 8 号 p. 1227-1235
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーFish school is an example of the familiar form of animal social behavior. However, the mechanism of the schooling behavior has not been clear yet. In our earlier papers, a mathematical model has been proposed for describing the behavior of fish school by using observation data from water tank experiments. Each fish regulates his movement, speed and direction, by the information which he receives from other fish around him. The mechanism of this information exchange has been demonstrated by carrying out computer simulation.
In this paper, a special attention is paid to analyzing the model for the schooling behavior of two fish. This model is represented by a system of nonlinear differential equations with five state variables. We investigate the relationship between a stable steady state of the system and a schooling behavior of fish swimming in the two-dimensional unbounded space. A result of analysis shows that the steady states of fish are given as a rectilinear motion and a circular motion. From an investigation of the local stability of the steady states it is found that the fish behavior model constructed by water tank experiment data is classified into four groups. Each group has different stable steady states. The most dominant group of the model has such a stable steady state as two fish swim along a linear trajectory and the leader corresponds to the fish with larger characteristic velocity.抄録全体を表示PDF形式でダウンロード (1272K) -
張 平, 山海 嘉之, 太田 道男1995 年 31 巻 8 号 p. 1236-1238
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーA hierarchical learning control using neural networks is proposed for adaptive control of nonlinear systems. Parameters of controller are adjusted by genetic algorithm in higher level learning and by dynamic back-propagation in lower level. Its effectiveness has been shown by computer simulations.抄録全体を表示PDF形式でダウンロード (374K) -
中野 和司, 内田 主幹, 江口 三代一, 中原 一夫1995 年 31 巻 8 号 p. 1239-1241
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーExtensions of the results of the model reference adaptive control to nonlinear systems have rarely been achieved. For this purpose, an alternative method of exact model matching (EMM), based on a kind of simulator called “dynamic network” which assigns any damping characteristics, is presented for a class of nonlinear mechanical systems. This approach is extended to the case of inverted pendulum system with no matching conditions.抄録全体を表示PDF形式でダウンロード (346K) -
山村 毅, 大洞 将嗣, 大西 昇, 杉江 昇1995 年 31 巻 8 号 p. 1242-1244
発行日: 1995/08/31
公開日: 2009/03/27
ジャーナル フリーWe simulate the process of interpreting Rubin's reversible figure as a human being does, based on some perceptual findings. Successful results were found for normal figure. However, further study is needed for the up-side-down case.抄録全体を表示PDF形式でダウンロード (339K) -
1995 年 31 巻 8 号 p. e1
発行日: 1995年
公開日: 2009/03/27
ジャーナル フリーPDF形式でダウンロード (112K)
- |<
- <
- 1
- >
- >|