巻号一覧
15 巻, 2 号
選択された号の論文の24件中1～24を表示しています
- |<
- <
- 1
- >
- >|
-
米澤 洋, 大川 不二夫1979 年 15 巻 2 号 p. 141-146
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーIn spite of the large number of designing methods for an adaptive observer proposed by the researchers of stability theorem, the study of an adaptive observer with observation noises has apparently not been published to date.
In this paper, a designing method for the discrete adaptive observer based on the Lyapunov's direct method is proposed. It is shown that the proposed adaptive algorithms are sufficiently valid where observation noises are present and that they are identical with the adaptive algorithms of a noise-free system.
Futhermore, are investigated the convergence speed and the steady state value of errors between the parameters of the observer and those of the system.抄録全体を表示PDF形式でダウンロード (620K) -
雛元 孝夫1979 年 15 巻 2 号 p. 147-152
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーThis paper presents methods to construct an observer and an adaptive observer for the bilinear systems described as follows:
x(k+1)=ui(k)Ax(k)+Bu(k)
y(k)=Cx(k),
where x(k) is an n×1 state vector, u(k) is an r×1 input vector, y(k) is an m×1 output vector, ui(k) is the i-th element of the vector u(k), and A, B, C are real matrices of proper dimensions. An example of this type of bilinear systems appears, for instance, in a heatexchange process. In this paper, it is supposed that
rank [CT, ATCT, …, (An-1)TCT]T=n.
The observer proposed here has an advantage that the estimated value of the state vector agrees with a true one in a finite number of steps and the adaptive observer also has the same property for the input with sufficient generality.抄録全体を表示PDF形式でダウンロード (702K) -
舟橋 康行, 安達 由洋1979 年 15 巻 2 号 p. 153-158
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーFrom the viewpoint of state estimation, we may regard bilinear systems as time-varying systems and apply to them the results with the observer of a linear time-varying system. However, since the input is in general not known a priori, it seems impossible to apply the theories of time-varying linear systems to bilinear systems. For the purpose of simple implementation, a new type of state observer is proposed in this paper.
The output of the bilinear system is one-time differentiable. Based on this fact and on the concept of Lyapunov stability, we propose a method for designing a state observer for bilinear systems, using information on the input, output and one-time deriverative of the output. This state observer has a specific feature of simple structure and the norm of the estimation error converges to zero asymptotically irrespective of the input. As a result, state observers are shown to be designed for a wider class of bilinear systems. Also the condition, under which a state observer without information on one-time derivative can be designed, is given obviously.抄録全体を表示PDF形式でダウンロード (730K) -
金井 喜美雄, 野口 哲郎1979 年 15 巻 2 号 p. 159-164
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーThe subject of modern control theory has received increased interest in recent years, partly because of its importance as an analytical tool for the study of advanced control technology. However, the main disadvantage of the modern control technique is that all the state variables and the parameters of a plant to be controlled are required to be accessible by measurement.
In this paper, two different adaptive observer schemes to improve the convergence charactaristics are developed to estimate the state variables and to identify all the parameters of a single-input single-output linear system where only the input and output can be measured. The computational algorithms are derived so as to minimize a quadratic performance. The approach is proved to be asymptotically stable and thus the convergence of the identification algorithm is ensured.
Numerical computations are done for the parameter identification and state estimation of a second order system. It is concluded that the proposed algorithms are simple and have a quick convergence.抄録全体を表示PDF形式でダウンロード (669K) -
稲垣 眞人, 舟橋 康行1979 年 15 巻 2 号 p. 165-170
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーThis paper considers the reachability of discrete bilinear systems described by x(t+1)=Ax(t)+Nx(t)u(t)+bu(t) t=0, 1, 2, …, where u is scalar control. This problem has a solution when the control is bounded or rank [N, b]=1.
In this paper there is no such limitation on the control. First, the properties of the set of states reachable in t steps from a nonzero initial state are investigated. Next, a sufficient condition for the existence of initial states, from which any state in Rn can be reached in n steps, is presented. As a result the control sequence that transfers the initial state to any desired state can be obtained in an explicit way.抄録全体を表示PDF形式でダウンロード (696K) -
杉本 末雄, 伊東 一文, 杉山 博1979 年 15 巻 2 号 p. 171-177
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーThe second order moment stability problems for second order linear stochastic systems with colored noise coefficients are studied. Namely, the proposed system equations are ÿ0+2ζν0y0 +[ν02+εx(t)]y0=0 and ÿ0+[2ζν0+εx(t)]y0+ ν02y0=0, where x(t) are the solution processes through the n-th order shaping filter with Gaussian white noise inputs.
After the joint Markov models are constructed by Ito's stochastic differential equations with coupling the coefficients processes as state variables, the exact second order moment equations regarding with y0(t)-processes are derived in iteration formulas by applying the Ito's differential rule.
Finally, under small ε assumption for the coefficients processes, the approximate necessary and sufficient conditions up to the order of ε3 are obtained via perturbation approach. It is shown that the conditions can be expressed by the power spectrum density functions of the x(t)-processes and only two spectrum point of x(t), 0 and 2ν0√1-ζ2 are related to the obtained stability conditions.抄録全体を表示PDF形式でダウンロード (740K) -
松尾 強, 長谷川 泰道1979 年 15 巻 2 号 p. 178-185
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーIn this paper, the general realization problem of discrete time systems including nonlinear, time-varying, and infinite dimensional systems is discussed.
The general realization problem is stated as follows.“Given a black-box whose input values and output values belong to arbitrary different sets U and Y respectively, find a“Mathematical Model”of the black-box which fits to the input-output data.”
Our“Mathematical Model”in this paper is a dynamical system. The definition of the general dynamical system was proposed in 1960, while the dynamical system theory was established in 1969 by R. E. Kalman only for linear, time-invariant, finite dimensional systems.3) Recently, T. Matsuo has established the realization theory of general dynamical syetems for continuous-time case.1)
We present the realization theorems of general discrete-time systems by rewriting Matsuo's theory. The special features of discrete time systems are as follows. 1) There exists only one concatenation monoid FS(N+, U), namely, a free monoid U* over the set U of input values. Note that several kinds of concatenation monoids exist in continuous-time case. 2) An Ω-module (X, Ψ) can be described by a U-action (X, F). 3) An. Ω-morphism is equivalent to a U-morphism.
By products of this realization theory, we have that: 1) An automaton is a special discrete-time dynamical system. 2) The condition of minimal control system which includes the minimal automaton is equivalent to the condition of multiple experimentablility.抄録全体を表示PDF形式でダウンロード (1214K) -
河野 通夫1979 年 15 巻 2 号 p. 186-192
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーThis paper considers the stabilization of the linear periodic discrete-time system by the use of a linear periodic state-variable feedback. Let the transition matrix of the closed-loop system be Y(t, s). Then the stability of the closed-loop system depends on the eigenvalues of Y(τ, 0), where τ is the period. It is shown that there exists a periodic state-variable feedback gain matrix such that the eigenvalues of Y(τ, 0) take pre-assigned values if and only if an open-loop system is completely reachable.抄録全体を表示PDF形式でダウンロード (891K) -
明石 一, 今井 弘之1979 年 15 巻 2 号 p. 193-198
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーThis paper considers the problem of designing an output deadbeat controller, using a state feedback such that the closed loop system is stable, in discrete time linear multivariable systems. The definitions and characterizations of a zero output stabilizable subspace and a zero output controllable subspace are given. It is shown that a stable output deadbeat controller can be designed if and only if the pair (A, B) is controllable modulo the zero output stabilizable subspace, where A is the transition matrix and B is the driving matrix. Finally, a procedure is given for designing the output deadbeat controller when it can be designed. The resulting controller is optimal in the sense that the controller settles the output in zero for any initial state in the minimum number of steps.抄録全体を表示PDF形式でダウンロード (755K) -
渡部 慶二, 伊藤 正美1979 年 15 巻 2 号 p. 199-204
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーGenerally, it is difficult to stabilize systems with long time delay in control and/or output by means of a conventional PID controller. A process-model control such as the Smith predictor has been utilized for controlling such systems. However, it is known that when the Smith predictor is applied to astatic controlled systems with time delay, the steady state error is not zero for additive disturbances, even through an integrator is introduced as a compensator. The same problem emerges in the process-model control proposed by Fuller.
In this paper, the steady state characteristics of a process-model control are investigated from the system-structural point of view. This study reveals that the steady state error is zero if and only if the state variable of the integrator is unobservable at the output of the process model. A process-model control which satisfies this condition is also presented.抄録全体を表示PDF形式でダウンロード (790K) -
添田 喬, 芳村 敏夫, 小西 克信1979 年 15 巻 2 号 p. 205-212
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーThis paper is concerned with a problem to identify the parameters of multi-variable bilinear systems where the output observation is contaminated by additive noise. Those systems may be considered to belong a class of time-varying linear systems with a known structure, and the model reference adaptive techniques are applicable to identify the parameters. Firstly, an asymptotically stable recursive identifier is constructed in a deterministic way by introducing the concept of strictly-positive systems for time-varying linear systems. Secondly, it is proved that the estimates of the parameters converge to the true values almost surely under some boundedness conditions.
This result is an extention of those obtained for linear systems, however, the proof of convergence given in this paper is more intuitive and simpler than the proof given for linear cases. A computational example for a second-order bilinear system is presented to illustrate the utility of the proposed identifier.抄録全体を表示PDF形式でダウンロード (1000K) -
田村 捷利, 坂田 伸弘, 岡村 和典1979 年 15 巻 2 号 p. 213-220
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーIn this paper, adaptive observers for a single-input, single-output linear unknown plant are considered. It is shown that the observers can be classified into six types according to the stability condition and the structure of the observer. A unified design method of the six types of adaptive observers is described after the preliminary lemmas, which are necessary in the design, are presented. The adaptive observers which have been reported in the articles fall into the four types in the above classification. The remaining two types are newly given in this paper as a result of the unified approach. The characteristics of these observes are summarized and the numerical simulation is given.抄録全体を表示PDF形式でダウンロード (973K) -
喜多村 直, 明石 一1979 年 15 巻 2 号 p. 221-227
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーThis paper deals with the problem of reducing the order of periodic linear time-invariant control systems which are described by input-output data with no noise.
One goal of the procedure given in this paper is to derive from input-output data the frequency transfer function matrix of the model by expanding it to the finite Kautz orthonormal function system. The other is to compose a completely controllable and observable model of the transfer function matrix by using the Gilbert's realization method. The order of expansion and the model dimension have a simple relation depending on the data, where they are identical in the case of a single input and/or a single output.
A uniform approximation criterion being is introduced, the relation between the dimension of the reduced model and the error between model and system is made definite.
As a practical example a dog systemic circulation is reduced to a two-dimensional model by the method above. This example will be useful in designing the adaptive control system of the cardiac assist pump being developed by the authors.抄録全体を表示PDF形式でダウンロード (864K) -
宮崎 孔友, 瀬崎 正博1979 年 15 巻 2 号 p. 228-234
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーThis paper describes a non-contact method for measuring the flatness of a mirror or a mirror-like surface by reflection of light beams from a group of line light sources arranged horizontally on a plane meeting at an angle of 45° with the horizontal plane. The light beams from these light sources are projected onto, in the first place, a liquid surface at rest, and in the next place, a mirror surface to be measured. With each projection the reflecting light beams are focused into an image on a screen separately provided, and by comparing both meanderings of the images of line light sources on the screen, the authors obtained the flatness of the mirror, the deviation between the mirror and the ideal geometrical plane. In tis paper, the authors propose the measuring principle, measuring errors and experimental results of the above method, and make it clear that the flatness of the mirror surface can be obtained mechanically at an arbitrary magnifying ratio by measuring only the meanderings of the images of light sources.抄録全体を表示PDF形式でダウンロード (2038K) -
相川 長彦, 金住 順二1979 年 15 巻 2 号 p. 235-240
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーFlow coefficient measured by an area-type flowmeter can be represented as a function of the Reynolds number through a series of experiments. From the experimental results obtained in this study, it was clarified that the characteristic curve can be divied into three ranges, that is, the viscous flow range, the intermediate flow range and the nonviscous flow range depending on the value of the Reynolds number.
At first, in the viscous flow range, laminar flow theory can be adoped suitably to the flow at the clearance between the taper-tube and float. Secondary, in the intermediate flow range, the values of the two sorts of coefficient, that is, one is that for geometry of the clearance between the taper-tube and the float and the other is that for the distribution of flow rate, were obtained as a function of the Reynolds number experimentally.
And then, it was observed that the float with sharp edge was useful for the measurement of viscous flow.
Furthermore, the nonviscous flow range seems to be a special case of the intermediate flow range because the nonviscous flow range corresponds to the case of an extremely, large Reynolds number in the intermediate flow range.
From the discussions on the value of α done above, the following experimental equations were derived.
Those equations are well representing the experimental results with a high degree of approximation.
α=α0(Re/Re0)1/2
for the viscous flow range
α=α0(1-K/√Re)(-1N/√Re)
for the intermediate flow range
where α; flow coefficient
α0; flow coefficient of nonviscous flow range
Re, Reθ; Roynalds number (Re0 is obtained through the experiment)
K; coefficient of rate distribution depending on Reynold number
N; coefficient geometrical condition depending on Reynolds number抄録全体を表示PDF形式でダウンロード (732K) -
山田 光, 苅屋 公明1979 年 15 巻 2 号 p. 241-245
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーThe aim of this paper is to propose a new method for the pulse height analysis in which a successive approximation type A-D converter (A. D. C.) is used. Because of its large differential non-linearity (D. N. L.), ordinarily the successive approximation type A. D. C. is not usable for the pulse height analysis. But this A. D. C. has a very short conversion time, so, an input pulse waveform is divided into many sample points. With these sample points, the height of a pulse peak is estimated by a suitable processing, namely, by a data processing through which the noise and D. N. L. are decreased. In this method the average amplitude ratio of an input pulse to the standard pulse is taken. Then this average is proportional to the real pulse height. By using this method, it is shown that the successive approximation type A. D. C. can be used for the pulse height analysis with a small D. N. L..抄録全体を表示PDF形式でダウンロード (1232K) -
本多 敏, 山崎 弘郎1979 年 15 巻 2 号 p. 246-251
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーIn the area of the fluidic measurement by the use of ultrasonic wave propagation, the existence of gas bubbles in liquid causes some trouble, due to scattering and absorption of the wave.
A new method for non-contact detection and measurement of bubbles in blood is developed utilizing the same phenomena.
In this paper, a new apparatus for the measurement is described. Its theoretical analysis on the attenuation of ultrasonic wave in two-phase fluid is also discussed. Two different implementation are tested, these are, (1) the measurement through the average attenuation of the transmitted wave, (2) the measurement through the RMS fluctuation of the wave power. As the results of the experiment, the theory was verified and the method was proved to be promising. The minimum detectable concentration of bubbles is 100ppm.
Our development work was aiming at the detection of small bubbles in human blood during the dialysis treatment. In this application the vantage of this method is fully demonstrated, which employs highly sensitive instruments not in direct touch with blood. Some results of field trial in a hospital are given. Also, a possible application to the industrial measurement is mentioned.抄録全体を表示PDF形式でダウンロード (768K) -
宮武 信春, 中村 豊1979 年 15 巻 2 号 p. 252-258
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーIn this paper, we compare the four evaluation methods to construct the multiattribute utility function based on the data of decision maker's preferences (A_??_A', A∼A', A_??_A). The first method is the Keeney's approach constructing an additive or multiplicative utility function based on the expected utility hypothesis. The second is the Probit analysis approach which estimates the weights of additive utility by maximum likelihood estimation method. The third is the Bayesian approach which estimates the weights of additive utility by Bayesian estimation method. The last is the discrimination function approach which estimates the weights of additive utility by the statistical correlation ratio maximization method. These methods are applied to the property value analysis.抄録全体を表示PDF形式でダウンロード (934K) -
岡沼 徹1979 年 15 巻 2 号 p. 259-266
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーThe real-time trajectory estimation of the lowest-flying earth satellite by a square root filter is discussed in this paper. A perigee altitude of this satellite, which is attained by firing an on-off capable solid rocket motor, is 100∼150km, therefore a real-time trajectory estimation is required for monitoring the altitude.
This filter is a non-linear square root filter which uses tracking radar data of an earth station and is designed to suppress to a minimum the divergence in filtering by a special device. The capability of this filter is examined by a precise simulation program.
As a result of the simulation, it is shown that the estimation error of perigee altitude is less than 1km and the computation time of this filter suits to a real-time operation.抄録全体を表示PDF形式でダウンロード (1059K) -
荒木 光彦, 森 武宏1979 年 15 巻 2 号 p. 267-268
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーPDF形式でダウンロード (227K) -
石島 辰太郎1979 年 15 巻 2 号 p. 269-270
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーPDF形式でダウンロード (255K) -
高原 康彦, 池庄司 英臣1979 年 15 巻 2 号 p. 271-272
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーPDF形式でダウンロード (312K) -
玉野 和保, 砂田 謙二, 北山 正文1979 年 15 巻 2 号 p. 273-274
発行日: 1979/04/30
公開日: 2009/03/27
ジャーナル フリーPDF形式でダウンロード (266K) -
1979 年 15 巻 2 号 p. 276
発行日: 1979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 (142K)
- |<
- <
- 1
- >
- >|