All issues
Volume 6 (1970)
- Issue 6 Pages 483-
- Issue 5 Pages 385-
- Issue 4 Pages 289-
- Issue 3 Pages 195-
- Issue 2 Pages 97-
- Issue 1 Pages 1-
Volume 6, Issue 6
Displaying 1-13 of 13 articles from this issue
- |<
- <
- 1
- >
- >|
-
Kunio YOSHIDA, Mitsuo TAKAOKA, Shozo IMAO, Isao MIWA1970 Volume 6 Issue 6 Pages 483-489
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSAs is well known, a DC chopper using a single thyristor is operated with the variable-frequency TRC (time ratio control) system. However, this system has the disadvantage that the control range in load voltage is limited. From this point of view, therefore, we propose the new method for DC chopper which utilizes a minimum triggering level mode performance of the thyristor. If our proposed method is used, a constant-frequency system can be operated with the square-wave control signal in a wide range of the load voltage.
In this paper, the operating principle and the steady state analysis of this system are described. In addition, the temperature compensation for this chopper is discussed with some experimental results. The chopper was applied to control the speed of a DC series motor. From the experimental data, it is clear that the chopper was operating stably to the inductive and counter-emf loads.View full abstractDownload PDF (1361K) -
Toshiyuki UENO1970 Volume 6 Issue 6 Pages 490-497
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper describes a purely inertial system whose attitude around its axis is controlled with air jet, and gives the results of the adaptive optimum control of the system. It is well known that the time optimum control of the purely inertial system can be accomplished by appling positive and negative maximum torques according to the parabolic switching curve in the phase plane.
In order to realize the desired switching curve, a nonlinear resistor, called the “varistor”, is used in the control system. Keeping the time optimum control loop closed, the small perturbation signal is applied to the system and the inertial moment of the controlled subject is measured by the correlation method. The measured moment is fed into the adaptive control system to determine the optimum switching curve in the phase plane automatically. By changing the moment, it is experimentally verified that the switching curve is optimally set.
The response of the moment measuring system to the sudden change of the moment is described. Also the effects of the test signal on the time optimum control and on the error of the measured moment are stated.View full abstractDownload PDF (962K) -
Yasushi TSUJIOKA, Hiroto SAEKI, Nobuyoshi TASAKA1970 Volume 6 Issue 6 Pages 498-507
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe stability of linear multivariable systems, consisting of identical channels and having antisymmetric cross-connections, is examined.
By considering that the structural features of multivariable systems have been classified into P, V and H-canonical structures, the transfer functions with complex coefficients are derived for each structure, and they are termed as the complex transfer function.
Further, it is shown that n-channel systems under consideration are described from the view point of stability in the form of series connection of the subsystems made up of one or two-channels.
In a composite system, consisting of two-channel systems, the stability of the entire system can be investigated by analyzing the stability of the individual subsystems.
It is also shown that the entire system can be extensively stabilized by attaching a two-channel system with antisymmetric cross-connection to any dynamic system.View full abstractDownload PDF (1199K) -
Tsuyoshi SAKANISHI, Takeo HORII1970 Volume 6 Issue 6 Pages 508-516
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn this paper the authors consider the identification problem as one of the essential functions of an adaptive control. The G(z)-method is proposed to be used for identifying the single input-and -output linear and nonlinear continuous systems, as well as the multivariable continuous systems.
If the pseudorandom signal input is given to a continuous system and its output is sampled synchronizingly, the system can be regarded as a sampled-data system with a zero order hold circuit and samplers. Obtaining the coefficients of the pulse transfer function of this equivalent sampled-data system, the parameters of the continuous system are calculated directly from the coefficients.
From the viewpoint of adaptive control, an identification error is mainly due to noise, time varying parameter (T.V.P.) and nonlinearity. Up to now the pseudorandom noise has been used to obtain the system impulse response topologically, so that some restrictions placed on the identification time present the T.V.P. error, especially in the multivariable systems.
Meanwhile the G(z) method has more flexibility for eliminating causes of the errors mentioned above. Also the apparatus of identification and data treatment become simple. The G(z) method is a useful real time identification method for the adaptive control systems.View full abstractDownload PDF (959K) -
Suguru ARIMOTO1970 Volume 6 Issue 6 Pages 517-522
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper is concerned with an information-theoretical approach to identification problems of a discrete-time linear system with constant coefficients. The system's input-output relation is described in the sequel by y(t)=t-1Σk=t-Nh(t-k)u(k)+z(t) (A1) or y(t)=nΣk=1[an+1-ky(t-k)+bn+1-ku(t-k)]+z(t) (A2) where h=(h(1), …, h(N))is an approxirnated finite sequence of the unknown impulse response and φ=(a1, …, an, b1, …, bn) is a collection of coefficients of the transfer function. The problem is to estimate a true value of h or φ by giving an appropriate sequence of test signals {u(t)} and observing the output sequence {y(t)} which is subject to the noise {z(t)}. Much attention is paid to selecting an optimal input sequence for estimation in the sense that it maximizes the mutual information under the constraint |u(t)|≤A.
In case of system (A1), it is shown that the best way is to select a sequence of binary input signals with as small correlations as possible. In case of system (A2), an analogous discussion would formally be possible, but the resulting scheme of selecting an optimal sequence of input signals is not feasible because the present value of input signals depends on the values of future outputs. To establish a feasible and reasonable method for selecting input signals, the concept of step-by-step mutual information is introduced and a step-by-step maximization algorithm is obtained.View full abstractDownload PDF (831K) -
Yoshimi TESHIGAWARA1970 Volume 6 Issue 6 Pages 523-530
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSLately, the traffic congestion in road networks has become a serious problem as cars increase. The enormous cumulative waiting time at signals is not only a loss of time but a cause of air pollution. A systematic signal control may decrease the loss.
Considering the signal control of traffic, there are three parameters, that is, the cycle time, the cycle split (ratio of green time to cycle time) and the offset (phase difference between two signals). Concerning the traffic networks, the offset control is the most effective and important of the three.
It is known that there is an optimum offset between two neighbouring signals which minimizes the waiting time of two-way traffic flow. However, if a traffic network system is very large and complicated, it is very difficult to find the combination of all offsets, that is, the offset pattern, which minimizes the total waiting time for the whole network. The purpose of this paper is to find the optimum offset pattern.
The author shows such traffic network systems can be optimized by using the technique of dynamic programming, if the waiting time of each pair of intersections is given as a function of offset. Moreover, he applies linear programming to this problem by approximating the waiting time by two linear functions and finds easily the optimum solutions of more complicated systems using a macroscopic model of traffic flow.
The optimum offset patterns of some network examples obtained by D.P. and L.P. are compared with the results of macro-model simulation. In applying the optimum control obtained by D.P., the total waiting time is decreased by 40 to 70 percent compared with that when there is no control (random offset), and by 10 to 50 percent compared with that when the simultaneous signal systems or the alternate signal systems, or the favoring systems are adopted. Similarly, the offset patterns obtained by L.P. decrease the total waiting time by 15 to 60 persent with no control and by 3 to 35 percent with control.View full abstractDownload PDF (927K) -
1970 Volume 6 Issue 6 Pages 530
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSDownload PDF (80K) -
Mituo KURAKAKE1970 Volume 6 Issue 6 Pages 531-536
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSIn many plants, we do not know all the component parts of the system whose state variables are to be estimated. The problem to be considered in this paper is the identification of unknown parameters and the estimation of state variables of a discrete-time noisy linear system with unknown elements in its transition matrix. The unknown elements are regarded as unknown parameters.
It is known that, for a linear system without unknown parameters, the optimal estimation process using the method of least-square fit can be replaced with a Kalman-Filter. The author suggests to utilize the performance index used in the method of least-square fit when the linear system has some unknown parameters to be identified. A parameter is identified when it gives a minimum of the performance index. The identification, therefore, is reduced to seek the minimum value of the perfomance index by changing the unknown parameters. Identifiability of unknown parameters is defined, and necessary and sufficient conditions for it are obtained, from the convergence of Kalman-Filter and from the equivalence of discrete systems.
In this method, the system with unknown parameters is treated always as a linear system, the Kalman-Filter to the system can be applied on the similar conditions as to a linear system without unknown parameters. Two examples prove the possibility of identification of unknown parameters.View full abstractDownload PDF (709K) -
Toshiyuki KITAMORI1970 Volume 6 Issue 6 Pages 537-545
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSAn adaptive process control system is developed making the ultimate sensitivity method, due to J.G. Ziegler and N.B. Nichols, on-line. The essential point for making it on-line is to keep the amplitude of the ultimate oscillation so small that the oscillation gives no appreciable disturbance at the controlled process output. The ultimate oscillation gives the required information about the controlled process dynamic characteristics, that is, the ultimate sensitivity and the ultimate period, according to which the parameters of the PID process controller can be adjusted. The ultimate sensitivity is the reciprocal of the process gain at the frequency where the phase-lag is 180°. The ultimate period is the reciprocal of the frequency. Ziegler and Nichols gave appropriate formulae relating the process controller parameters, i.e., the proportional gain, the integral time and the derivative time to the ultimate sensitivity and the ultimate period. The adaptive control consists of the ultimate oscillation amplitude control, the ultimate period measurement, and the controller parameter adjustment.
Simulation studies are carried out on an analog computer, giving satisfactory results. The ultimate oscillation is generated and the amplitude of which is controlled stably at a sufficiently low level. The proportional gain, the integral time and the derivative time of the PID controller are adjusted proportionally to the ultimate sensitivity and the ultimate period measured, according to the formulae. The step response of the control system approaches a response curve with very short rise-time and with a suitable damping factor.
The dynamics of the ultimate oscillation amplitude control are analysed in the appendix, for the design of stable amplitude control. Also in the appendix, the dynamic behaviour in the adjustment of the integral time and the derivative time is explained, giving the stability condition for the system.View full abstractDownload PDF (2357K) -
Masataro NISHIMURA, Shinzo KITAMURA1970 Volume 6 Issue 6 Pages 546-553
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSCalculation of the integrated squared error (ISE) for linear systems with time lag is studied by using the Lyapunov functional method of Repin, and an error in his method is corrected. The method described here makes it possible to obtain the exact values of the ISE through the solutions of certain linear differential equations and algebraic equations.
The analytical forms of the ISE for the first-order system KcKe-Ls/1+sT and the second-order system Kc(1+1/TIs)Ke-Ls/1+sT are obtained. In numerical examples, the relationship between the ISE and the system parameters is given, which is, for reference, compared with the parameter setting proposed by Ziegler and Nichols.
It is also shown that the instability condition for the first-order system, which can be obtained from this Lyapunov functional, is identical with the necessary and sufficient condition for the stability of the system.View full abstractDownload PDF (788K) -
Kageo AKIZUKI, Katsuhiko SHIRAI, Tetsushi SASAGAWA1970 Volume 6 Issue 6 Pages 554-561
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThis paper is concerned with the problem of the influence of Gaussian white noise on a selfoscillating control system.
Two possible methods of analysis are applied to a nonlinear feedback control system containing a saturating element. One is the averaging method and the other is the statistical equivalent linearization method which has been used widely for nonlinear control systems with random inputs or disturbances.
The results obtained by these methods are quite different.
Namely, according to the statistical equivalent linearization method, the variance will jump and the limit cycle will cease to exist.
However, by the averaging method, neither the jump phenomenon in the variance nor the cessation of limit cycle is predicted. An analog computer simulation confirms the conclusion of the averaging method.
The difference between the two methods lies in the domain of linearization.
In the averaging method, the system is linearized around the limit cycle, but in the statistical equivalent linearization method, the system is linearized roughly in the larger domain.
Though the both are approximation techniques, the averaging method is more suitable for the analysis of this type of system.View full abstractDownload PDF (754K) -
Hidemine OHNO, Kazuo SAGARA1970 Volume 6 Issue 6 Pages 562-569
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSA method of analysis of multi-variable control systems by digital computer is discussed in this paper. A large scale system can be decomposed into a suitable number of subsystems which are represented by signal flow graphs. In this method, the frequency characteristics of each subsystem is calculated by matrix elimination in the first place. Then, a matrix equation of a composed system is made up by rearranging elements of transfer matrices of the subsystems. From this equation, the transfer matrix of the higher class system can be evaluated by matrix elimination, just the same as in the signal flow graph reduction.
This method has much computational advantage to save the time for computation and to use the memory space effectively, as compared with the reduction without the system-decomposition technique. If we introduced a hierarchical system-decomposition or grouping, the computational advantage would be increased remarkably. Especially, a case study, in which parameters of the system are changed, can be carried out with little additional time for computation through an adequate system-decomposition. For this purpose the routine for reduction is proposed in this paper.
In the computer program developed by the authors, the input data can be made up directly in the tabular form from a block-connection diagram and signal flow graphs of a given system.
As an example, a steam temperature control system of reheat boiler is presented to demonstrate the computational advantage of this method.View full abstractDownload PDF (1075K) -
Tadashi NAGATA, Masao MUKAIDONO, Ryoichi MUROI1970 Volume 6 Issue 6 Pages 570-578
Published: 1970
Released on J-STAGE: March 27, 2009
JOURNAL FREE ACCESSThe manipulator control with a digital computer has been studied mainly in the exploration of the outer space and the deep ocean. On the other hand, the idea of applying a manipulator with some intelligence to the industries has existed for some years. The typical ones belonging to the industry are Versatran and Unimate. These and other so-called industrial robots have been put to practical use. Industrial robots of this sort, having only simple functions such as recording and regenerating, seem to have very limited applications.
In this paper, a computer control system of the industrial robot with conversational mode is described. The industrial robot is controlled by means of a digital computer through effective communications with an operator. This control system is similar to MANTRAN developed in M.I.T. for supervisory control of a remote manipulator, and has the following features so as to find suitable applications in industry.
(1) the operator can teach the robot various kinds of jobs through conversations with a computer,
(2) the robot grows wise as it gathers experiences, and the operator has to teach less accordingly.
(3) When the robot cannot make a decision, the operator is ready to make a suitable decision for the situation, and
(4) it is not necessary to prepare programs for every kind of situation at first, and the robot can construct new job command freely and easily by combining job programs accumulated already.View full abstractDownload PDF (1854K)
- |<
- <
- 1
- >
- >|