All issues
Volume 10, Issue 1
Displaying 1-18 of 18 articles from this issue
- |<
- <
- 1
- >
- >|
-
Pyong Sik PAK, Yutaka SUZUKI, Katsuhiko FUJII1974 Volume 10 Issue 1 Pages 1-5
Published: February 28, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn classical control systems, integral-type controllers are widely used to cope with process parameter variations and other perturbations or disturbances. In this paper a method based on the regulator theory to synthesize an optimal multivariable servo-system incorporating the integral-type controllers is described. The principle of this method is similar to the previous one1), but is different in that the integrated values of the tracking errors are introduced as the new process state variables. The advantages of this method are as follows:
(1) Optimal control is easily obtainable.
(2) The resulting control system has feedback and feedforward pass and has optimal tracking properties. Even if process parameters vary from their nominal values, steady state errors do not arise.
In an actual process, state variables are often inaccessible. For such a process the control system can be constructed by utilizing an observer. It is shown in the example that the input-output characteristics of a control system using the observer are less influenced by the process parameter variations than are those of a state feedback system.View full abstractDownload PDF (579K) -
Seiichi MIYAZAKI, Tokuji OTA1974 Volume 10 Issue 1 Pages 6-14
Published: February 28, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSSince 1968, Chiyodakako Kensetsu, Fujidenki and Fujitsu have been researching for digital control of chemical processes.
In this project, two DDC systems have been developed, (the CPU of one system is FACOM 270-25 and the CPU of the other is FACOM R), and for an example of advanced control of a DDC system, the optimal control of a distillation column which uses state variables has been investigated. The summary of the paper is as follows.
(1) Result of experiments of static characteristics and dynamic characteristics of distillation column, and corresponding results of digital computer simulation. They have clear nonlinearity.
(2) Important equations of mathematical simulation model and linearized system equation for the application of optimal control theory.
(3) Method of calculation of optimal feedback gain and accuracy of solution of Riccati's equation.
(4) Ratio of the weights in performance index of control variables to the weights of controlled variables varies control action, but there was no convenient method to fix the ratio.
(5) Direction of control variable (reflux flow in this case) of optimal control was different from the direction of conventional top-bottom control.
(6) As the control has proportional action only, application of this control is limited.
(7) All the process variables which appear in the system equation must be fed back.
Figures show static temperature distribution and transient responses of experiments or digital computer simulation.View full abstractDownload PDF (1141K) -
Yoshikazu NISHIKAWA, Nobuo SANNOMIYA, Takeo OJIKA1974 Volume 10 Issue 1 Pages 15-21
Published: February 28, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper proposes a coordination algorithm for multilevel control of a large-scale dynamical system. The system considered consists of weakly interconnected nonlinear subsystems, and the performance index is quadratic in states and controls.
According to the variational principle, the optimal control is given by solving a nonlinear two-point boundary-value problem, of which an analytical solution is generally impossible.The present technique is to solve the overall problem, first by solving decomposed problems of the subsystems, and second by coordinating interactions among the subsystems. Since each subsystem problem is a linear two-point boundary-value problem, it is relatively easy to solve. The present idea of coordination is to adjust the interaction variables directly by an iteration without using the conventional Lagrange multiplier. A sufficient condition for convergence of the iteration algorithm is presented in the paper.
The algorithm is computationally simple and the convergence is quite rapid for the problem of weakly coupled systems with small nonlinearities. The effectiveness of the method is illustrated in an example.View full abstractDownload PDF (824K) -
Takashi SOEDA, Toshio YOSHIMURA, Tatsuki ISHII1974 Volume 10 Issue 1 Pages 22-27
Published: February 28, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper is concerned with a problem of deriving an approximate filter for noisy nonlinear, discrete-time systems with unknown time-varying parameters.
Firstly, a linearized filter is obtained by applying a Taylor series expansion to the nonlinearities of the system where the unknown parameters are assumed to be time-invariant.
Secondly, this filter is tested whether or not the approximate innovation sequence is likely to have come from the calculated distribution for it. When the hypothesis test is rejected, the calculated covariance matrix for the approximate innovation sequence is modified so that the approximate innovation sequence lies on the boundary of the acceptable region.
Numerical results of digital simulation for simple nonlinear systems indicate that the proposed filter gives better filtering performance than the filter presented by Tarn et. al.View full abstractDownload PDF (704K) -
Hayatoshi SAYAMA, Yoshimasa KAMEYAMA, Hirotaka NAKAYAMA, Yoshikazu SAW ...1974 Volume 10 Issue 1 Pages 28-35
Published: February 28, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe purpose of this paper is to describe the multiplier method for finding the solution of constrained nonlinear programming problems. The method is based on transforming a given constrained problem into a sequence of unconstrained problems by defining the multiplier function.
The algorithm proposed is to determine unconstrained minima of the multiplier function for fixed parameters and some Lagrange multipliers. The Lagrange multipliers are then varied by a simple correction and the multiplier function is minimized again. It is proved that the multiplier method is locally convergent to the saddle points of the multiplier function at the rate of a geometric progression by proceeding in this fashion.
The features of the multiplier, function are that it is a class of the generalized Lagrangian and is twice continuously differentiable at the boundary of the feasible region. The method deals with non-convex programming problems and determines the Lagrange multipliers as well as the solution of the original problem. The numerical difficulties encountered with existing penalty methods are avoided. Numerical examples are presented to illustrate the typical convergence characteristics of the method.View full abstractDownload PDF (974K) -
Approximated Calculation Method of the Transition Matrix Using Páde's Approximate Function of exp(x)Kajiro WATANABE1974 Volume 10 Issue 1 Pages 36-44
Published: February 28, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSA transition expression of a system is often used when the engineer faces the problems of digital simulation, digital estimation and control.
In this case, tedious calculation procedures and a great amount of calculation time have been needed for the transformation of the system expressions, namely, the transformation from the state variable equation to the transition equation.
This paper tries to solve these problems by the following method, that is, to compute the matrix function P=exp(Aτ), which is called the transition matrix, where A is the system matrix of the state variable equation. Páde's approximate function
P=I+Aτ(I-0.5Aτ)-1,
which can be calculated easily and has some properties similar to those of the exponetial function, is used instead of the finite Taylor series expansion method of exp(Aτ).
As a result of using this method, it can be said that this method not only shortens the calculation time, but also ensures the numerical stability of the numerical solution of the linear stable differential equation, which is one of the important properties of numerical problems. And furthermore, the system matrix A of the state variable equation can be easily calculated, inversely, from the transition matrix P, as follows
A=2/τ(P+I)-1(P-I),
which has important applications to digital estimation.View full abstractDownload PDF (1027K) -
Toshihiro KOBAYASHI, Yoshikazu SAWARAGI1974 Volume 10 Issue 1 Pages 45-55
Published: February 28, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSDistributed parameter differential games and sensitivity synthesis of optimal strategies for a game with terminal constraints are discussed in this paper.
In the first part of the paper, minimax problems are analyzed for functionals defined on real Hilbert spaces, where an existence theorem for saddle points is derived. As applications of the above theory, distributed parameter differential games with distributed controls, boundary controls or pointwise controls are considered. In each case, the necessary and sufficient conditions for existence of saddle points are derived in the form of variational inequalities.
Secondly the game with terminal constraints is investigated, where the playability property plays an important role. The sensitivity synthesis of optimal strategies for the problem is developed to make up for the undesirable effects caused by the system parameter variations. In this sensitivity game, the central role is played by the combined system, which consists of a model equation and its parameter sensitivity equation. To ensure the existence of optimal strategies, the playability property of the combined system is examined. By giving simple examples, the effectiveness of the present sensitivity synthesis is illustrated.View full abstractDownload PDF (1291K) -
Shigeki HATAYAMA1974 Volume 10 Issue 1 Pages 56-62
Published: February 28, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSGeneral delay-differential control systems dependent on both the previous history of the state and of the control are modelled. The optimal control problem for such systems with the cost functional, the state constraint condition and the time-varying target set is formulated. This formulation generalizes the optimal control problem examined by M.N. Oguztoreli. An existence theorem of an optimal policy in the class of absolutely continuous initial state functions in the C space and measurable control functions in the L2 space is proved. Further, the uniqueness of an optimal control in the case of linearr delay-differential systems is considered.View full abstractDownload PDF (832K) -
Kiyotaka SHIMIZU, Yuichiro ANZAI1974 Volume 10 Issue 1 Pages 63-70
Published: February 28, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSOptimization problems for hierarchical systems are formulated such that local systems have their own independent objectives within some restrictions imposed by a central system. In those systems, the central system governs local ones by means of-the resource allocation to them and optimizes its own objective by suitably allocating resources to them. On the other hand, each local system optimizes its own objective consuming the given resources. An optimization algorithm for those problems is studied based on multiparametric linear programming. Some practical examples are given to indicate the usefulness of our formulation.View full abstractDownload PDF (1000K) -
Kenko UCHIDA, Kinji MORI, Etsujiro SHIMEMURA1974 Volume 10 Issue 1 Pages 71-77
Published: February 28, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper deals with the stochastic control of Linear systems with nonclassical hierachical information structures. Optimal control problems and optimum information processing problems are discussed under this information structure. Person-by-person optimality and optimality are defined, and a sufficient condition for person-by-person optimality is derived. A condition, under which the person-by-person optimal control and the optimal control coincide is also given. Using this condition it is shown that the linear-quadratic-Gaussian problem has solutions which are linear functions of the optimal estimates. As optimum information processing problems, the optimum adjustment of the information delays and the optimum timing of the observations are discussed.
To clarify the underlying ideas and to avoid inessential complexity of the formulas, only the two-controllers case will be considered in the main part of the paper. The extension to the multiple-controllers case is conceptually straightforward.View full abstractDownload PDF (918K) -
Norio BABA, Yoshikazu SAWARAGI1974 Volume 10 Issue 1 Pages 78-85
Published: February 28, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper discusses the learning behaviours of variable-structure Stochastic Automata under stationary random environment.
A new reinforcement scheme (TNP) of the reward-penalty type which can ensure ε-optimality under all stationary random environments is proposed. It is proved, using Semi-Martingale Inequality and complex manipulations, that the TNP scheme can ensure ε-optimality. Moreover two reinforcement schemes (Lr-1 and T1) which have been contrived are discussed from that point of view.
The TNP scheme is superior to Lr-1 and T1 in the following respects ((1), (2)):
(1) The Lr-1 scheme is the reward-inaction type.
(2) The T1 scheme can ensure ε-optimality only under certain conditions.
Computer simulation results also indicate that the TNP scheme accomplishes the most effective learning behaviour.View full abstractDownload PDF (794K) -
Osamu KATAI, Sousuke IWAI1974 Volume 10 Issue 1 Pages 86-92
Published: February 28, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn a practical testing hypothesis, sample data is often used to construct testing rules for input information. It is empirically known that if we have small samples, we must compress or reduce the input information to obtain satisfactory results.
The paper discusses the above relation from a Bayesian theoretical point of view.
First, we take notice of the structural properties of the information and its compression form the probability theoretical point of view. Consequently, the effect of information compression on Bayesian testing rules and the relation to Bayes risk are presented.
As it is difficult to treat the risk theoretically, we introduce an information theoretical measure closely related to it. The measure enables us to evaluate the effect of information compression in a Bayesian testing hypothesis.
The structural properties of information and the above evaluation determine the rules for optimum information compression.
Finally, the effect of sample size on the optimum compression is made clear.View full abstractDownload PDF (985K) -
Michio NAKANO, Kensuke HASEGAWA1974 Volume 10 Issue 1 Pages 93-99
Published: February 28, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe author formerly proposed a rotating magnetic field type step motor. This step motor holds the follwing advantages:
(i) When one input pulse arrives, the rotating angle of the rotor that corresponds to this input pulse can be controlled by a simple eletric circuit, and therefore this angle can easily be varied in a wide range.
(ii) The heat loss at the time of lightload, compared with the conventional step motor, is extremely small. Especially, at standstill and no load(when output torque is zero), the electric current in the step motor is exciting current only, so the heat loss of this motor is negligible.
(iii) By the use of a universal wound-rotor induction motor, a high power step motor can easily be obtained.
An alternating magnetic field type step motor possessing the same advantages as the rotating magnetic type step motor is now proposed for the first time. The principle of the alternating type step motor is as follows. At first, an alternating magnetic field is made on the stator of the motor which has two phase windings. This magnetic field is shifted by the digital control circuit as each input pulse arrives, and then the rotor is attracted.
In this paper, the control method of the new step motor is discussed, and the torque is calculated by tensor analysis. By using this control circuit, the induction motor was driven as the step motor and the basic experiment was conducted. The results of this experiment were successful.View full abstractDownload PDF (779K) -
Eiji TOBA, Hiromichi TOYOTA, Oichi ATODA1974 Volume 10 Issue 1 Pages 100-106
Published: February 28, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSA special CA thermo-pile with high sensitivity and fine frequency characteristics, of which the gain is flat to 10Hz, is developed as a sensor for measurement of temperature fluctuations.
The irregular temperature fluctuations produced around a heater are convected by air flow and detected by the two sensors provided along the flow. Using the distance between two sensors and a cross-correlation function of the two signals, the flow velocity is calculated.
The main features of this method for, the measurement of the air flow velocity are as follows:
(1) The method using temperature fluctuations is more advantageous than tracer injection methods, becausee it does not contami nate an object.
(2) In the region of 0.5m/s∼10m/s, measurement of the flow velocity is possible if the velocity is calibrated with a standard flow meter.
(3) This method requires some time interval for averaging times, since the correlation method is used to obtain the flow velocity.
(4) The relative error of this measurement is as small as ±5% when the product (vl) of the distance between two sensors (l) and the flow velocity (v) is greater than 0.12 for l=10cm, and 1.40 for l=30cm.
Moreover, when vl is smaller than the abovementioned values, the errors are systematic Thus this method is still useful with compensation.
The velocity measurement by this method is applicable in various fields of flow.View full abstractDownload PDF (835K) -
Takayoshi TOTANI, Hideo KATO1974 Volume 10 Issue 1 Pages 107-113
Published: February 28, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe effect of mixing andd diffusion of turbulence upon the accuracy of the flow velocity measurement by means of the cross-correlation technique is discussed in this paper. The transport process of a random fluctuation of physical properties in turbulent flow is treated in the domain of frequency by assuming that the process is described by the diffusion equation, and the difference beween the delay time of this process and the distance/velocity lag is estimated. It is shown that the diffusion effect causes the delay time at the peak position of the cross-correlation function to be shorter than the distance/velocity lag, and that it is fully possible to keep the measurement error due to this cause within 0.5% by properly selecting the dimension and the disposition of the detectors.
The other source of the static error is due to the inequality of the dynamic characteristics of the two detectors. This error can be reduced by utilizing the high frequency component of the signals.
The theoretical considerations are compared with experimental results on an air pipe flow with velocity fluctuations.View full abstractDownload PDF (841K) -
Tadanao OHISHI, Yoshimasa SAKURAI, Shoichi SEINO1974 Volume 10 Issue 1 Pages 114-121
Published: February 28, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper describes a method for analog measurement of interference fringe displacement and evaluation of its systematic error.
The interference fringe of a two-beam interferometer is vibrated in a small amplitude by frequency modulation of a laser, which is applied either as a light source of the interferometer or by an oscillating glass plate inserted in one of the arms of the interferometer. The fringe is detected by a photoelectric system containing a reversible counter, a count-voltage converter and a low-pass filter. The reversible counter and the count-voltage converter convert the momentarily changing position of the vibrating fringe into voltage. The voltage is filtered to remove the modulation component and the conversion distortion. Then the output voltage is analogous to the fringe displacement.
The systematic error of this measuring method is calculated numerically on some wave form models of the modulation signal. The amount of the error depends on the wave form of the modulation and its amplitude. For a triangle wave, a condition is found under which measurement is carried out without any amount of error. But in general, this measurement does include an error, the magnitude of which is less than λ/30.View full abstractDownload PDF (898K) -
Iwao AMASAKI, Tsukasa SUNOSE, Ichiro KATO1974 Volume 10 Issue 1 Pages 122-126
Published: February 28, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn this paper, the authors describe the physical mechanism relating to a control characteristic of the Johnsen-Rahbek effect and controlled by a third electode.
It is found that the operation of the third electrode displays some new physical phenomena.
The results obtained are as follws:
1. The mechanical force is controlled by the third electrode potential.
2. The control characteristic is of V-shape and the mechanical force produced depends on the potential difference between a main voltage and a control voltge.
3. When the main voltage is equall to the control voltage, the mechanical force is produced.
4. The peeling off time of the three electrode type is one order shorter than that of the two electrode type.View full abstractDownload PDF (649K) -
Sadao FUJIMURA, Shigeki SAGAYAMA1974 Volume 10 Issue 1 Pages 127-133
Published: February 28, 1974
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSA method for segmentation of quantized patterns from a standpoint of connectivity is proposed. A pattern consisting of black points connected to each other is separated as a unit. These points are labeled during a scan of a picture. By an algorithm shown in this paper the same label is given to the points that join together. After one scan of the picture, a pattern is separated from the others by making use of the labels given to the points. The use of this method makes it also possible to count patterns. The number of labels used indicates how many patterns are contained in the picture. An improved rather noise-free method is also propoosed, in which a pattern is regarded as compsed of segments instead of points. The method enables us to pick up information while scanning for the segmentaion, which is necessary to extract everal kinds of features from each pattern with ease. These are global features, such as a circumscribed rectangle, center of gravity and axes of symmetry, and partial ones, such as corner points, loops and end points. The information is available for discriminating patterns or for adapting the segmentation algorithm to objects.View full abstractDownload PDF (2007K)
- |<
- <
- 1
- >
- >|