計測自動制御学会論文集
Online ISSN : 1883-8189
Print ISSN : 0453-4654
ISSN-L : 0453-4654
27 巻, 4 号
選択された号の論文の20件中1~20を表示しています
  • 清水 優史, 香川 利春
    1991 年 27 巻 4 号 p. 381-385
    発行日: 1991/04/30
    公開日: 2009/03/27
    ジャーナル フリー
    To know the characteristics of spherical float area-type flowmeter, a series of experiments covering a wide range of float weight and Reynolds number have been conducted. From the experimental results, the following empirical formula relating the non-dimentional flow-rate, float weight and minimum flow cross-sectional area was obtained:Q=173(ArW0.91d/83)2.6/(ArW0.91d/83)1.5+1 where Ar is nondimensional minimum cross sectional area and Wd is nondimensional effective float weight.
    This formula can express all the experimental results, except for very low Ar and Wd case, within 5% error.
    Based on the empirical formula, the effect of working condition change was discussed and the correcting method indicated in JIS was shown not to be appropriate. An accurate method correcting working condition change was proposed.
  • 川田 剛之, 大谷 敦久, 日下 迢
    1991 年 27 巻 4 号 p. 386-393
    発行日: 1991/04/30
    公開日: 2009/03/27
    ジャーナル フリー
    The reflected near infrared radiance of rugged terrain as measured from space, includes strong atmospheric and topographic effects. The removal of such effects from satellite image data is very important in the analysis of remote mountainous forest areas. However, the radiative transfer process is much more complex for rugged terrain than for flat terrain, because the complicated local topography, such as elevation, slope and aspect, must be taken into account.
    The summary of this study is as follows: 1) In this study we have developed a new analytical correction method for rugged terrain image data which allows the use of the more general Minnaert's law for the bidirectional reflection at the ground surface. In our correction model, single reflection by a target slope and multiple reflection by the mean back ground as well as diffuse reflection by the atmosphere were taken into account. 2) The validity and effectiveness of the proposed method were sucessfully demonstrated to the MOS-1 MESSR image data. 3) We found that the reflection by rugged terrain vegetation (mostly decidious forest) follows Minnaert's law, rather than Lambert's law. The appropriate values of the Minnaert constant k and mean ground albedo A for the study site were k=0.59 and A=0.56±0.06 in band 3 and k=0.63 and A=0.79±0.08 in band 4.
  • 渡辺 嘉二郎, 小林 一行, 小山 明美
    1991 年 27 巻 4 号 p. 394-401
    発行日: 1991/04/30
    公開日: 2009/03/27
    ジャーナル フリー
    The correspondence problem to identify the trajectories of the several moving points, e.g., the centers of moving objects in a sequence of images with a static scene is one of the important problems in the dynamic scene analysis. Assuming that the movements of the points could be described by the equation of motion, the correspondence problem for several non-rigid objects which seem to move arbitorarily in the scene is considered.
    We formulate the correspondence problem as a prediction problem by searching pairs of real points and predicted points which have the minimal distance. The prediction points are obtained by fitting the trajectories of moving objects to a 2-dimensional 2nd order auto-regressive model with assist of the rules about the movements when the movements of the points cannot be correctly described by the prediction model. An algorithm to find trajectories of points in a image is presented.
    We demonstrate the validity of this approach by experiment to track the random walks of several persons in a real scene.
  • 広井 和男, 戸苅 吉孝
    1991 年 27 巻 4 号 p. 402-408
    発行日: 1991/04/30
    公開日: 2009/03/27
    ジャーナル フリー
    This paper describes the two degrees of freedom PID (2 DOF PID) algorithm with compound filter which has newly been developed.
    Though there are many papers published regarding to various kind of 2 DOF PID algorithm such as set point filter type and separation of P, I and D, they still have problems when practical application.
    The new method here proposes combination of a compound filter and a conventional PI algorithm. This can be advantageous to industrial practical application as realizing the complete 2 DOF PID algorithm by simple programming and the highly accurate nonlinear PID control performance simply, easily and freely.
    This paper points out that the varieties of 2 DOF PID algorithms depend on the structure of filter in the algorithm. The procedure of compound filter is now proposed as it may realize the new 2 DOF PID algorithm.
    Furthermore, we confirm the effect of the proposed algorithm by simulation.
  • 平沼 賢次, 河野 通夫
    1991 年 27 巻 4 号 p. 409-414
    発行日: 1991/04/30
    公開日: 2009/03/27
    ジャーナル フリー
    This paper considers numerical algorithms for pole assignment. First, the algorithm for a single-input system is given. This is based on ideas from the QR method, and numerically reliable. Second, for a multi-input system (A, B) without assuming a special structure, a numerical algorithm is proposed for the computation of the state feedback matrix such that the closed-loop system is cyclic. The algorithm is based on the Householder transformations. Thus, we change the multi-input problem into a single-input problem and then apply the previous algorithm. Minimis and Paige suggested the algorithm which is based on the QR method, where the starting point is a block Hessenberg form. Since our method does not assume a special structure, it is simpler than the method of Minimis and Paige. A numerical example is also given.
  • 近藤 良, 原 辰次, 伊藤 敏也
    1991 年 27 巻 4 号 p. 415-421
    発行日: 1991/04/30
    公開日: 2009/03/27
    ジャーナル フリー
    A unified state-space solution for continuous and discrete time H control problems is presented using bilinear transformations.
    First, a computation algorithm for solving both continuous and discrete time Riccati equations is proposed by introducing the Hamiltonian matrix for the discrete time system, which has the same form as that for continuous time system. The J-spectral factorizations of transfer functions in continuous and discrete time cases are developed by completely parallel way.
    Finally, we present the unified and simple parameterization of all solutions of continuous and discrete time H problems in the state-space setting, and it is shown that the solvability conditions are same for both continuous and discrete time cases.
  • 羅 正華, 坂和 愛幸
    1991 年 27 巻 4 号 p. 422-428
    発行日: 1991/04/30
    公開日: 2009/03/27
    ジャーナル フリー
    This paper deals with an experimental study on coupled bending and torsional vibrations of a flexible arm, which occur when a flexible arm carries a large rigid load. The purpose of this paper is to show through experiments that the tip position of the arm and the coupled bending and torsional vibrations can be controlled simultaneously by manipulating a driving motor. To do this, a linear dynamic model, which consists of two decoupled partial differential equations and a set of coupled ordinary differential equations, is first derived. Then, this complicated model is formulated as an abstract equation and a feedback control law is designed on the basis of the model. The experimental results confirm the accuracy of the derived dynamic model and the validity of the feedback control law.
  • 櫛田 直規, 西村 行雄, 米澤 洋
    1991 年 27 巻 4 号 p. 429-435
    発行日: 1991/04/30
    公開日: 2009/03/27
    ジャーナル フリー
    In this paper, the authors describe a time optimal control system which is needed for quick course changing to specified direction, for example, on the path of a narrow channel. Hence, the authors design the system as an adaptive time optimal control one assuming that manoeuvrability is the 3rd order linear system which consists of both steering mechanism (integral element restrained state) with fixed parameters and ship dynamics (disturbed first lag plus integral element) with unknown parameters, and is also maintained the control ability regardless of parameter variation given by outer cause, as wind, current, wave and stain of the ship, and by inner cause, as load or velocity.
    The proposed control system has following some features:(i) Reduction of computing time for solving the time optimal problem by new predictive strategy method using a simple algebraical solution.
    (ii) Addition of PID controller for overcoming unstability caused by the parameter/construction miss match around neighborhood of the origin.
    (iii) Adaptive identification (by Landau's algorism) of controller parameters using the suitable mathematical model to maintain the control ability.
    Finaly, the authors show the results of both the digital simulation and the sea trial, and confirm good effectivness by the proposed system.
  • 吉田 和弘, 広瀬 茂男
    1991 年 27 巻 4 号 p. 436-443
    発行日: 1991/04/30
    公開日: 2009/03/27
    ジャーナル フリー
    The 3D vision is indispensable for realization of an intelligent robot. The stereo vision has been considered as a typical candidate for 3D visions, but it is not yet practically used. The bottleneck in realizing the stereo vision was a corresponding problem which is to correctly identify the same viewing point for each camera. The paper proposes a new stereo vision system which solves the corresponding problem with multiple arrayed cameras system. The proposed vision system, named “MAC vision”, has the following characteristics, i) real-time measurement function of 3D vision, ii) simplicity and inexpensiveness of signal processing system, iii) parallel processing of many kinds of feature points. The paper also discusses the several problems to realize MAC vision and the experimental results of the constructed model of MAC vision with five cameras. From the experiments it is shown that the MAC vision can measure indoor and outdoor scenes with the speed of 16 frames/s, as to the frame including 1, 800 points of 3D coordinates measurement.
  • 小野木 克明, 川上 茂人, 竹村 学, 西村 義行
    1991 年 27 巻 4 号 p. 444-451
    発行日: 1991/04/30
    公開日: 2009/03/27
    ジャーナル フリー
    A problem of seeking feasible solutions under a large number of constraints is decomposed into a set of interconnected subproblems. If two subproblems share same variables, equality constraints arise between the subproblems. These equality constraints represent the interconnections of subproblems. This paper first shows that such a problem can further be transformed into a two-level optimization problem composed of small-scale optimization subproblems and one coordination problem. Auxiliary coordination variables are introduced to coordinate the common decision variables included in two or more subproblems. It second proposes a method of seeking feasible solutions based on cooperative in distributed problem solving. Feasible solutions are sought through the minimization with respect to the decision variables in each subproblem of the first level and through the minimization with respect to the coordination variables in the second level. If many feasible solutions exist, the coordination in the second level causes a multimodal problem. The paper also discusses a problem of separating peaks from multimodal function. The feasible solution seeking method proposed here seems to easily manage knowledge necessary to problem solving, to decrease effects of curse of dimensionality, and to have potential possibilities of parallel processing.
  • 伊藤 聡, 志水 清孝
    1991 年 27 巻 4 号 p. 452-457
    発行日: 1991/04/30
    公開日: 2009/03/27
    ジャーナル フリー
    This paper deals with an infinitely constrained optimization problem-a nonlinear programming problem with an infinite number of constraints. This problem can be regarded as a satisfaction optimization problem, in which an objective function is to be minimized subject to the satisfaction conditions that plural performance criteria should be kept below the prescribed permissible level even under the worst situation. As computational methods for the problems of this sort, there are the relaxation algorithm and the nondifferentiable optimization algorithm.
    In this paper, we propose the dual quasi-Newton algorithm as a direct method for infinitely constrained optimization problems. First of all, this class of problems is reformulated as an optimization problem with an infinite-dimensional inequality constraint by introducing an abstract operator; then we apply the quasi-Newton method to the abstract optimization problem. This means that an infinite number of constraints are simultaneously taken into account. The subsidiary problems for direction-finding are still infinitely constrained, but their dual problems are nonnegatively constrained quadratic programming problems in a function space and they can be easily solved by clipping-off techniques without any relaxation or applying nondifferentiable optimization algorithms.
  • 神余 浩夫, 竹垣 盛一
    1991 年 27 巻 4 号 p. 458-465
    発行日: 1991/04/30
    公開日: 2009/03/27
    ジャーナル フリー
    In this paper, we propose the COordinative Decentralized control system Architecture: CODA, as a fault-tolerant industrial plant control system. The system can be reconfigured to adapt to its conditions and environment by the coordination of its controllers. It consists of symmetycally coupling autonomous controllers, and has neither supervisory controller nor reserved one. These controllers communicate with each other and decide their own functions. In order to make the communicating time short enough to keep a realtime constraints, they communicate and cooperate locally. Hence, the global system must repeat these local cooperations till getting to a coordinating state; we call this behavior the coordinative dynamics. For example, when a controller is broken, normal controllers backup its functions locally and gracefully degradate under overloaded condition. And this changing is influences to far controllers gradually. To gurantee such a coordination, we designed Coordination Protocol as the rules of cooperations among the controllers. This Protocol consists of three function layers; the Primitive Communication Layers, the Distributive Scheduling Layer and the Autonomous Scheduling Layer. Using the Coordination Protocol, it is easy to apply other control systems. We analyzed the ability of this Protocol, and made its prototype to confirm the abilability of this framework.
  • 村越 英樹, 土肥 康孝
    1991 年 27 巻 4 号 p. 466-473
    発行日: 1991/04/30
    公開日: 2009/03/27
    ジャーナル フリー
    This paper proposes a high speed programmable controller which executes sequence control specified by Petri Net. The programmable controller can perform logical operations for checking and firing transitions in parallel using a special hardware unit, named Fire Unit. And it decreases scan time for picking up input signals by means of interrupt mechanism. As Fire Unit can check whether a transition is fireable or not in 2 memory cycles, the programmable controller can execute Petri Net control in six memory cycles. Assuming that it is implemented by memories of 100 nano second memory cycle, it performs Petri Net control in 600ns. It is more than 100 times faster than conventional programmable controllers. We designed Fire Unit in gate level logic, and showed that it can be implemented as an ASIC memory in the current VLSI technology.
  • 北川 洋一, 金谷 典武, 林 昭博, 南 茂夫
    1991 年 27 巻 4 号 p. 474-476
    発行日: 1991/04/30
    公開日: 2009/03/27
    ジャーナル フリー
    The effects of particle size distribution on the output of the laser-scattering fiber-optic particle size monitor are discussed. The relations between the particle size distribution and the error in the particle size decision are evaluated.
  • 渡 正博
    1991 年 27 巻 4 号 p. 477-479
    発行日: 1991/04/30
    公開日: 2009/03/27
    ジャーナル フリー
    This paper proposes a new algorithm of two-color radiation pyrometry, in which spectral emissivity is a function of wavelength. Test results using heated stainless steel sheets showed that the temperature error of the new method due to the emissivity variation hysteresis is less than those of conventional methods.
  • 小野 敏郎, 孫 建新
    1991 年 27 巻 4 号 p. 480-482
    発行日: 1991/04/30
    公開日: 2009/03/27
    ジャーナル フリー
    In this paper, we propose a new method for identification of linear systems. Applying the new method, it is possible that the identified dynamics is independent of the dynamics of the measuring instruments used to obtain the necessary data.
  • 大久保 元晶
    1991 年 27 巻 4 号 p. 483-485
    発行日: 1991/04/30
    公開日: 2009/03/27
    ジャーナル フリー
    This note presents a model-matching problem of the controller with the reduced-order observer. We obtain the closed-loop transfer function matrix Φ which is a simple affine function of the parameter matrix Q.
  • 大塚 尚久
    1991 年 27 巻 4 号 p. 486-488
    発行日: 1991/04/30
    公開日: 2009/03/27
    ジャーナル フリー
    A simultaneous block triangular decoupling and disturbance rejection problem with state feedback for linear multivariable systems is studied in the framework of geometric approach. Necessary and sufficient conditions for this problem to be solvable are presented. Further, an illustrating example is given.
  • 渡辺 啓史, 森田 矢次郎
    1991 年 27 巻 4 号 p. 489-491
    発行日: 1991/04/30
    公開日: 2009/03/27
    ジャーナル フリー
    We introduce topology-like structure into the set of strongly connected subtournaments, and inspect their structural relations. As a result, it is shown that the remoteness of cycles is characterised by the inclusion family (or pedigree) of cycles, and how different substructures appear in one strongly connected tournament.
  • H制御問題からのアプローチ
    Goujun SHI, Yun ZHOU, Chengwu YANG, 大須賀 公一, 小野 敏郎
    1991 年 27 巻 4 号 p. 492-493
    発行日: 1991/04/30
    公開日: 2009/03/27
    ジャーナル フリー
feedback
Top