All issues
Volume 10 (1974)
- Issue 6 Pages 641-
- Issue 5 Pages 521-
- Issue 4 Pages 389-
- Issue 3 Pages 263-
- Issue 2 Pages 137-
- Issue 1 Pages 1-
Volume 10, Issue 2
Displaying 1-18 of 18 articles from this issue
- |<
- <
- 1
- >
- >|
-
Hideaki KANOH1974 Volume 10 Issue 2 Pages 137-143
Published: April 30, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe aim of the present work is to study the static and the dynamic behaviors of a lumped parameter heat exchanger controlled by manipulation of the flow rate. It is shown that a PI-feedback control system has a unique steady state and that a P-control system is always stable, but that a PI-control system has an unstable region in which the limit cycle is caused by the nonlinearity of the manipulation of the flow rate rather than by valve saturation. The amplitude and the period of the limit cycle can be determined by approximate calculation. Experimental results verify the existence of the limit cycle and agree well with the theoretical calculation.View full abstractDownload PDF (698K) -
Toshihiko OHSAWA1974 Volume 10 Issue 2 Pages 144-152
Published: April 30, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSA new method is presented for preparing a gas mixture with precise composition and greatly changing its consumption.
This method is distinguished from conventional methods in the following aspects. A new technique was adopted in order to achieve precise control over the constituents of the gas mixture. A new ‘constant differential pressure control method’ is used for keeping the flow quantity under control. And a widely used surge absorber is employed to provide a continuous supply in accordance with changes in consumption.
The components of the gas mixture are completely mixed in the piping by a new type of mixer. They are described in detail.
The experimental results show that the maximum dispersion of the composition in a three-component system is less than 1%.
The difference between the data of the analysis is less than the inherent scatter of the analysis of a laboratory gas chromatograph.
These results show that the flow control system and the line mixer are entirely satisfactory.View full abstractDownload PDF (3032K) -
Osamu KATAI, Takashi IMANAGA, Sousuke IWAI1974 Volume 10 Issue 2 Pages 153-159
Published: April 30, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSStatistical approaches for quantization of patterns are discussed together with its asymptotic properties.
To discuss the problem quantitatively, we introduce an information measure closely related to the recognition rate of a Bayesian pattern recognizer.
Using this measure, we evaluate the information loss resulting from quantizing patterns as the amount of decrease of the measure.
We adopt the valuation as the performance index of quantization and consider optimum methods in the sense of the valuation for various kinds of quantization problems.
Experimental computer simulations are done in several cases and the relation to the Bayesian recognition rate is shown.
Further, to examine properties of these methods, we notice asymptotical properties in regard to the number of quantizing levels.
Consequently, their validity and the relation to the usual method based on the mean square error criterion are made clear.View full abstractDownload PDF (835K) -
Eizo KURODA, Katsuhiko FUJII1974 Volume 10 Issue 2 Pages 160-166
Published: April 30, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper presents a method to determine the time-optimal control of a linear system with a certain state inequality constraint. The method is fully concerned with the maximum principle. In this method we consider two trajectories, at least in principle, simultaneously. One of them starts from the initial state of the system and is computed in the forward direction of time, and the other starts from the target state and is computed backward. These trajectories are iteratively corrected so that they meet together, in an optimal fashion, on the boundary of the state admissible region. The optimality is tested on the basis of the geometric properties of the state attainability set that is particularly defined. In this method one does not have to explicitly consider the discontinuity of the adjoint variable at the junction time, and, further, the junction points are easily located. The computational procedure is given in detail and a numerical example is also shown.View full abstractDownload PDF (932K) -
Shin NYUDO, Akira HIGUCHI, Takeo SASAJIMA1974 Volume 10 Issue 2 Pages 167-173
Published: April 30, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSA direct method is presented in this paper for controlling the mechanical power generated by the prime mover without any type of energy conversion. A new power control system consisting of a clutch and a power spring is investigated The clutch is operated successfully by an ON·OFF signal, whose duty ratio is controlled to obtain the desired average. It is shown from analytical and experimental results that the proposed power control system has not only quick response, but very high efficiency. Further the system construction is simple enough to be implemented practically.
In the experimental study, the considerably high transfer efficiency, 75∼95%, has been achieved over a wide range of speed change. These indicate the practical usefulness of the system. An especially advantageous use of the proposed system will be made for the case where the start and stop signals are repeated with high frequency as in an electric train or motorcar.
It is also possible to realize regenerative braking, if the induction motor is taken as the prime mover.View full abstractDownload PDF (868K) -
Kunihiko ICHIKAWA1974 Volume 10 Issue 2 Pages 174-178
Published: April 30, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSA method using SUMT is proposed in this paper as a numerical solution method for a minimum time control problem. The characteristics of the paper presented are that the SUMT loop used for satisfying the terminal constraint and the other SUMT loop used for satifying the constraint for the magnitude of control are imbedded in one SUMT loop by regarding the solving method as a mixed interior point-exterior point algorithm and sequentially approaching the optimal solution of the final time and control function simultaneously by leaving the final time as an unknown parameter to beoptimized.View full abstractDownload PDF (578K) -
on Static CharacteristicsYoshio SUZUKI1974 Volume 10 Issue 2 Pages 179-186
Published: April 30, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper describes the static characteristics of a pneumatic micrometer with an ejector type nozzle. This type nozzle is a unit which consists of the nozzle, the output port and the supply port set up at a very near position from the nozzzle.
From experimental results, it was found regarding static output characteristics that the ouput flow rates and the output pressures became negative when the nozzle-flapper gaps became large, and regarding transfer characteristics, that the linearity was better than on that of the thickness comparator type pneumatic micrometer with a single nozzle.
A theoretical analysis was presented for the low supply pressure case, assuming that the pressure at the jet impinging part rose because of the momentum change of the jet impinging on the nozzle and the flapper. Then it was evident that a parameter λ0, which related to the pressure at the jet impinging part, explained the character of the ejector type nozzle and equaled approximately 0.36 when the design was done well.
The theoretical results showed a good agreement with the experimental results. The author gave a criterion for the design of an ejector type nozzle. This nozzle will be profitable as a pneumatic sensor, particulary when we need a large output flow rate for low loads, and it has wide applicability.View full abstractDownload PDF (925K) -
Kojiro HAGINO, Tomonao HAYASHI1974 Volume 10 Issue 2 Pages 187-193
Published: April 30, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe attitude of a satellite is very important not only for scientific observations but also for the house keeping, and attitude control of satellites is desirable. In order to accomplish the attitude control, measurements and determination of the attitude are necessary.
In this paper, the attitude determination of satellites by means of real time data processing is discussed.
Random errors are generally introduced to the output data of the sensors due to the noise in the electronic circuits and the random variation of the physical quantities measured, such as time and spatial variations of horizon. In the determination of the attitude, the choice of data processing in addition to the improvement of the acuracy of each sensor is necessary.
In this paper, attitude control using the geomagnetic field is assumed, and a filtering technique is applied to the attitude determination, which offers the basis of planning the attitude control. A method of attitude determination using a horizon sensor and a digital sun sensor is discussed, and the equations for the attitude determination are developed. The errors which should be expected in the measurements by each sensor are also discussed. By use of these results, digital simulations are made, and the accuracy and the time required in the attitude determination are studied. A sequential least squares method neglecting the dynamics of the attitude from the filtering technique is also discussed by simulations. The examples of the simulation show that 1∼4 degrees divergence in the attitude angle obtained from raw data can be improved to 0.1∼1 degree by both methods. When the filtering is applied and the estimated values of parameters in dynamics deviate largely from the true values, the error increases very greatly. From these facts, it is expected that the sequential least squares method is applicable in practice. It is confirmed that the attitude can be determined by both methods in one orbital period by utilizing all the data on the attitude in the orbit.View full abstractDownload PDF (822K) -
Yoshiaki TADOKORO, Tatsuo HIGUCHI, Takeshi ANAYAMA1974 Volume 10 Issue 2 Pages 194-199
Published: April 30, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe Fourier transform is used in many fields as an efficient method for signal analysis. But itss real-time processing has not come into wide use, because the necessary instruments are too expensive and not handy. This paper describes a method for the simple construction of a realtime Fourier analyzer and its application to magnetic torque curves.
The real-time Fourier analyzer is constructed by two basic components, a Fourier coefficient calculator and a memory element. The calculating circuit for the Fourier coefficient can be made easily by a newly developed circuit using a single operational amplifier which allows both positive and negative multiplication and addition. A magnetic element is used as the analogue memory element. The magnetic memory element is easily made from a magnetic core and has semi-permanent memory time. Practically, the real-time Fourier analyzer for the calculation of the Fourier coefficient with a scheme of 12 ordinates is constructed by these components. The characteristics of this analyzer are as follows. The zero drift is ±0.1%/hour on the full scale. The minimum amplitude of the analyzable input signal is about 0.02 in ratio to the maximum amplitude.
As an example for its application, magnetic torque curves are analyzed in real time. The analyzed results are almost identical with the results calculated by a general purpose computer in off-line. This simply constructed real-time analyzer will be suitable for the analysis of signals which do not have a very high frequency component, for example, magnetic torque curves, electrocardiograms, seismic data, etc.View full abstractDownload PDF (744K) -
Toshihiro KOBAYASHI, Yoshikazu SAWARAGI1974 Volume 10 Issue 2 Pages 200-210
Published: April 30, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper gives some considerations on the sensitivity of N-player, nonzero-sum distributed parameter differential games.
First, necessary conditions are derived for the three types of strategies (the Nash equilibrium, the minimax and the noninferior). The nominally equivalent property and/or the performance index value are investigated.
Secondly the optimal feedback strategies are defined, which consist of the output feedbacks and the prefilters and are nominally equivalent to the open-loop strategy. A general formula is derived to examine the sensitivity of the performance index, and the performance index sensitivity is indicated to be affected by how the optimal feedback strategies are chosen.
Finally other N-player, nonzero-sum games for sensitivity reduction are considered in order to synthesize the optimal feedback strategies. Several examples are given to illustrate this procedure.View full abstractDownload PDF (1310K) -
Nobuo OKUDA, Masahiko SEKIGUCHI1974 Volume 10 Issue 2 Pages 211-216
Published: April 30, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper describes relations between frequency measurement time and the stable operation of a pulse-drive servo system which is mainly used for servo drives in numerical control systems.
The pulse-drive servo has two main characteristics; first, it has high reliability, because it has only one noncontact angular position detector removing mechanical faults; and second, it has high stability, because its control circuits are constructed only with logical and switching elements, eliminating the temperature, drifts of amplifiers.
The velocity feedback signal, necessary for stable operation of the servo system, is obtained by measuring frequencies of feedback pulses. So the servo loop contains a time delay which is equivalent to the frequency measurement time and which causes unstable operation if the design is not done carefully.
Two typical constructions of the pulse-drive servo are proposed here; one is PDS-1, which obtains the difference between the position error signal and the velocity feedback signal using a pulse counter; and the other is PDS-2, which obtains the difference using an adder circuit.
It is shown that the characteristic equation of the servo system can be expressed by modified z transforms of transfer functions containing frequency measurement time as a parameter. Relations between frequency measurement time and stable operation of the servo are evaluated by root locus plots of the characteristic equation. And the results are compared with experimental data. The conclusion is drawn that PDS-2 is better for practical use because of its excellent stable operation.View full abstractDownload PDF (628K) -
Seiji KOBAYASHI1974 Volume 10 Issue 2 Pages 217-222
Published: April 30, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSNew methods are presented for the design of model-reference adaptive control systems. The motivation is in obtaining an adaptive system with faster adapting speed than the conventional systems. The methods of constructing adaptive loops have been given by many authors, but the difference between the outputs of the plant and the model is commonly used as the adaptive criterion. Therefore, the adapting speed has its limit due to the difference of initial conditions.
This obstacle is overcome by using adaptive criteria which do not depend on initial conditions. Two methods proposed here have criteria of that kind. One uses the adjoint system of the model, and the other is based on the substitution of the input and output of the plant to the differential equation of the model.
Adaptive criteria based on these ideas, the relations of these methods to the conventional one, and the constructions of adaptive loops by Lyapunov's second method are given. Finally the possibility for the improvement of the adaptive speed is shown through the simulation results of an example.View full abstractDownload PDF (731K) -
Kenko UCHIDA, Etsujiro SHIMEMURA1974 Volume 10 Issue 2 Pages 223-229
Published: April 30, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe optimal control problem of stochastic linear systems with time-delay is discussed. The problem is formulated as that of a coupled system of a lumped parameter sub-system and a distributed parameter sub-system which corresponds to a group of time-delay elements. This formulation clarifies that this problem can be treated as the Markovian control problem. In the framework of a class of stochastic Markovian control problem, a sufficient condition for the optimal control is obtained. Applying this condition, the optimal control for a quadratic criterion is derived.
The optimal control law can be realized by the cascade of a coupled type state estimator and a linear feedback. In other words, the optimal control is given in a form of a linear combination of state estimates x(t), ξ(t, ·) of both sub-systems. The optimal feedback gains are the solution of the Riccati type ordinary and partial differential equations, and coincide with that of the deterministic case.View full abstractDownload PDF (830K) -
Tokuji OKADA1974 Volume 10 Issue 2 Pages 230-237
Published: April 30, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper introduces a new method to recognize three-dimensional patterns by grasping an object with fingers. An experiment based on this method is shown at the end.
The hardware to apply this method has three fingers. The first finger, which corresponds to the thumb, has only one degree of freedom, and the second and the third, which correspond to the index and middle fingers respectively, have three degrees. Each pivot joint can be driven independently. Each supporting surface between pivots has ON-OFF tactile sense. Each torque of the joint is measured as the force sense.
The pattern recognition of an object is performed by analyzing the distributed pattern of the sensors in contact and the bending forms of the fingers when an object is appropriately grasped with the tactile and force control method which is provided to recognize the pattern.
The results of the recognition are used to grasp an object properly and to determine a control policy directly oriented to a task.View full abstractDownload PDF (1509K) -
Minoru MAEDA, Takashi SUZUKI, Toyomi OHTA1974 Volume 10 Issue 2 Pages 238-244
Published: April 30, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIt is well known that the optimization design for a linear system with quadratic performance index may be attained through the solution of the Riccati equation. In practical systems, it sometimes becomes very laborious and difficult to work out the solution, especially when the capacities of computers are limited. Kokotovic's method of singular perturbation may be considered to be suitable for these complex systems. The authors have extended this method to include higher order perturbation terms in order to improve the accuracy and proposed a mixed order perturbation method, considering the fact that the higher order terms do not necessarily contribute to the purpose.View full abstractDownload PDF (632K) -
Toshi TAKAMORI, Masatada YONEMOCHI1974 Volume 10 Issue 2 Pages 245-252
Published: April 30, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn use of pneumatic pipe lines in fluidics circuit systems for the purpose of information transmission, it is necessary to determine the suitable dimension of the pipe lines.
In this study, the dimension of the transmission pipe line giving a good matching to an information source and a load is determined from the power transmission coefficient of a fundamental transmission system composed from an information source with an internal resistance, a pipe line and a load resistance. The power transmission coefficient is expressed by a nondimensional frequency and three nondimensional parameters formed from the dimension of the pipe line and the values of the internal and load resistances. The effects of the nondimensional parameters on the power attenuation calculated from the power transmission coefficient are analytically investigated and experimentally examined.
Considering the result of these examinations, a method of design for a pneumatic transmission line is presented. It is finally explained with an example that the method of design is very practical.View full abstractDownload PDF (868K) -
Katsunori SHIDA1974 Volume 10 Issue 2 Pages 253-258
Published: April 30, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSTo eliminate a low frequency noise contained in a step DC signal, usually a lowpass filter is used, but when the frequency of the noise is very low, a great deal of time is required to get the final value of the step DC signal. The author has tried turning his attention in these circumstances to the rise-time interval of the DC signal. That is, the author has replaced the resistor ofthe ordinary RC onestage filter with a photoconductor-lamp composite element whose resistance can be controlled by a lamp current. The time-constant of this filter is made very small at the beginning of the DC step rise, and is increased proportionally to the time elapsed from the beginning. After a short time, the time-cnstant is made very large and is fixed. This procedure is triggered by the rise of the DC signal so the author calls this filter a “trigger filter”.
As a result of this, it has been found, theoretically and experimentally, that the output of this filter rises to within a given allowable range faster than the output of an ordinary RC one-stage low-pass filter. For example, when the noise is sine wave, the time required for the trigger filter to reach its response within the range of ±36% of the final value of the step response is less than that required for the ordinary RC filter.View full abstractDownload PDF (731K) -
Tsutomu YOKOZUKA, Eiji BABA1974 Volume 10 Issue 2 Pages 259-260
Published: April 30, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSDownload PDF (261K)
- |<
- <
- 1
- >
- >|