計測自動制御学会論文集
Online ISSN : 1883-8189
Print ISSN : 0453-4654
ISSN-L : 0453-4654
38 巻, 1 号
選択された号の論文の16件中1~16を表示しています
  • 宮川 透, 米澤 義道, 伊東 一典, 橋本 昌巳
    2002 年 38 巻 1 号 p. 1-8
    発行日: 2002/01/31
    公開日: 2009/03/27
    ジャーナル フリー
    In this paper, a new pen-based device without special writing plane is proposed. We tried to reproduce handwritten pattern using 3D pen acceleration and angular velocity measured by accelerometers and gyroscopes respectively. Pen positions are calculated with twice integration of pen acceleration considering pen rotation. The condition of the pen rotation is measured by pen angular velocity. In the process of twice integration, small noise and output drift of the sensors cause reproduction error. Pen velocity signals are rectified to remove the reproduction errors. Though coordinate system of reproduced pattern is slanted according to the initial pen slant, reproduced pattern can be transformed to the pattern on horizontal plane using gravitational acceleration. The experimental results show the possibility of a new pen-based interface system without special writing plane.
  • 山田 実, 徐 粒, 斉藤 制海
    2002 年 38 巻 1 号 p. 9-17
    発行日: 2002/01/31
    公開日: 2009/03/27
    ジャーナル フリー
    Some improvements have been proposed for a Bose's 2D (2 dimensional) stability test by revealing symmetric properties of the polynomials and the resultant in the test. We first show that Schüssler's 1D stability criterion can be tested with the complexity of computation in largely reduced as it involves only certain real variable polynomials with degrees not exceeding half of their previous counterparts. Based on Schüssler's 1D stability criterion, Bose has developed a resultant-based stability test for 2D systems. Using above result and the property that the resultant is Belt-inversive, we give an improved version of the Bose's 2D stability test. The proposed 2D stability test can be fulfilled more efficiently by totally algebraic algorithm and can be implemented more easily and systematically into a 2D CAD system. Non-trivial examples have also been shown.
  • 山下 裕, 大西 規雄, 西谷 紘一
    2002 年 38 巻 1 号 p. 18-25
    発行日: 2002/01/31
    公開日: 2009/03/27
    ジャーナル フリー
    In this paper, we propose a new method of exact tracking control for multi-input multi-output nonlinear systems in which the relative degree for an output is not well defined at the origin. First, the system is transformed into a singularly perturbed system using a time-scale transformation and coordinate transformations. A composite control law is adopted to attain exact tracking control for the singularly perturbed system. However, it is difficult to solve simultaneous partial differential equations for the composite control law. We obtain an approximate solution of the PDEs via the ε-correction method. The proposed tracking control law is applied to the ball and plate system, which is a typical example having no relative degree at the origin.
  • 川嶋 葵, 小林 一行, 渡辺 嘉二郎, 沼田 仲穂
    2002 年 38 巻 1 号 p. 26-34
    発行日: 2002/01/31
    公開日: 2009/03/27
    ジャーナル フリー
    This paper describes the modeling of a man-automobile system based on the a psychological man-machine model. The driver who must adapt to the frequently changing external environment in driving is receiving the psychological stress. Among variety of stimulations in driving, he selects some of them which may give him the injury and threat and/or to which he must pay effort to resolve, and percepts them as the noxious stressors. And, he begins to take action called coping to reduce the noxious factors in the stressor. This is an influential hypothesis of the psychology. If the driver drives the automobile based on this hypothesis, he keeps injury and threat away, releases himself from the vigilant mental situation and drives the automobile in the safe manner.
    Under the hypothesis above, we built a driver-automobile model and formulated the minimum-stress driving problem. The driving mode derived from the solution of the problem, is similar to that of human actual driving as well as the driving control structure is same as that of the conventional pre-view control well-known in the automobile engineering field. The simulation results and an unmanned vehicle controlled by the proposed formulation, shows the safe driving as man drive.
  • 楊 子江, 宮崎 大輔
    2002 年 38 巻 1 号 p. 35-44
    発行日: 2002/01/31
    公開日: 2009/03/27
    ジャーナル フリー
    Since a voltage-controlled magnetic levitation system is open-loop unstable and strongly nonlinear associated with the electromechanical dynamics, it is quite challenging to apply an adaptive controller for position tracking problem mainly due to its poor transient performance. This paper proposes a combined robust and adaptive approach to a popular magnetic levitation system. The overall controller is designed through a backstepping manner by combining both the adaptive control and robust control techniques. The input-to-state-stability is ensured by the robust nonlinear damping terms whereas the ultimate control error bounds are made sufficiently small by the adaptive laws. The combined approach proposed here helps to overcome some well-known practical problems such as high-gain feedback of the robust controller, and poor transient performance of the adaptive controller. Experimental results are included to show the excellent position tracking performance of the designed control system.
  • 最小到達時間問題として
    小柳 孝巳
    2002 年 38 巻 1 号 p. 45-51
    発行日: 2002/01/31
    公開日: 2009/03/27
    ジャーナル フリー
    The radiation relay is used for the measurment and control where a ordinary type level meter is not available. It is adequate for the control of the levels of melted metal, liquid in high pressure tank and etc… In this system the controlled object is placed beween a radiation source and detector. As the objct with uniform thickness covers the window of detector from one side to the other with uniform speed, the amout of incident radiation is reduced. At the first time when the output of detector reaches the given threshold, a relay acts and the motion of the object is stopped. Haw to design the system obtaining the given value of the mean stopping position and minimize the standard deviation of it is presented in this paper.
  • 川畑 良尚, 川上 知之, 川畑 隆夫
    2002 年 38 巻 1 号 p. 52-60
    発行日: 2002/01/31
    公開日: 2009/03/27
    ジャーナル フリー
    A new state-equation of the induction motor will be presented. The state-equation of induction motors is indispensable in design of vector control. In the vector control, high performance is realized by decoupling control. In case of the conventional state-equation, physical operation of the induction motor is difficult to understand. This is because, in the conventional state-equation, the physical meaning of the coupled terms are not easily discerned since the branches of the summing point don't have voltage dimension.
    In order to design the current controller, the transfer function of the motor after the decoupling control is used. If the conventional state-equation is used, this transfer function becomes the second-order including the rotor resistance. Therefore, it becomes difficult to design the current controller and the fluctuation of the transfer function occurs by temperature change of the rotor resistance.
    To overcome these problems, a new state-equation in matrix form which enables the physical meaning of the coupling terms to be easily discernable, is proposed. In this equation, all branches of the summing point have voltage dimension, and thus the equation enables easy understanding of the physical operation of the machine. Since the matrix theory is used, the existence of coupling between the magnetizing-axis and torque-axis becomes clear, and it is easy to understand the coupling between the stator circuit and the rotor circuit.
    Moreover, in the new state-equation after the decoupling control, the transfer function becomes the first-order without the rotor resistance. Therefore, it becomes easy to design the control system for decoupling.
    In addition, experimental results using the designed current controller are given.
  • 化学反応器への応用
    関 宏也, 大山 敏, 小河 守正
    2002 年 38 巻 1 号 p. 61-66
    発行日: 2002/01/31
    公開日: 2009/03/27
    ジャーナル フリー
    A nonlinear model predictive control based on successive linearization approach is developed. A nonlinear process model is linearized along its trajectory and an infinite-time horizon linear optimal regulator with integrator is implemented in a receding horizon fashion. Simulation examples for highly nonlinear chemical reactors which exhibit static input or output multiplicities are given to demonstrate the capability of the controller. The developed control algorithm has been successfully applied to a grade transition operation in an industrial polymerization reactor.
  • 鈴木 文泰, 堀 洋一
    2002 年 38 巻 1 号 p. 67-72
    発行日: 2002/01/31
    公開日: 2009/03/27
    ジャーナル フリー
    In this paper, the general structure of anti-windup control system based on left coprime factorization is proposed. First, the command responses of anti-windup control system are investigated using transfer function description, and the condition for avoiding windup phenomena is derived. The responses can be formulated easily by handling the input saturation as its equivalent input disturbance. Next, the conditions are explained using state space description. This analysis shows that the proposed system has very simple structure. For feedback controller designed by Youla parameterization, the proposed method can obtain the proper estimated value by state observer. For feedforward controller designed by right coprime factorization, the proposed anti-windup controller has the limitation model in the state feedback loop.
    The effectiveness of the proposed method is shown by some simulation and experimental results.
  • Hiroki YAMANASHI, Ikuya HOSHINO, Yoshihide OKAMURA, Haruo TANAKA
    2002 年 38 巻 1 号 p. 73-78
    発行日: 2002/01/31
    公開日: 2009/03/27
    ジャーナル フリー
    This paper describes the application result of the observer-based design method for the loop control of bullblock drawing machines. The salient features of this design method are that the disturbance observer is constructed as a constitutive control loop without a conventional PI control loop and the system design is accomplished in a more consistent way than the method which uses the disturbance observer in addition to the conventional control.
  • 伊藤 聡, 西垣 智啓, 川崎 晴久
    2002 年 38 巻 1 号 p. 79-86
    発行日: 2002/01/31
    公開日: 2009/03/27
    ジャーナル フリー
    In this paper, we consider standing posture control of biped system. When standing on a slope, a human adjusts his posture, i.e., ankle joint angles adjust to the gradient of the slope. In this case, the resultant stationary posture is not determined until the slope angle is given to the system. To solve this problem, we propose a new posture control law, where the standing posture is adjusted based on the ground reaction forces in such a way that the entire weight is placed evenly on the feet. This control law consists of PD control and force feedback control, whose stationary posture possesses local stability in adequate feedback gain. By applying this control law to a simple robot system, we examine its efficiency. In this experiment, we show the robot behavior is adaptive to the slope angle or external force generated by wind.
  • 西川 郁子, 西出 恵介
    2002 年 38 巻 1 号 p. 87-96
    発行日: 2002/01/31
    公開日: 2009/03/27
    ジャーナル フリー
    Multiple Coding Genetic Algorithm (MCGA) is proposed as a multi-objective GA method, in which each individual specializes on one of many objectives and has a different decoding according each objective. The proposed method is applied to 2-objective knapsack problems and 2-objective flowshop scheduling problem, and its performance is compared with the ordinal parallel selection. In the application to knapsack problems, the data sets with different correlation coefficients are generated in a systematic way to control the distribution of a Pareto set. Quantitative evaluations of the obtained solutions are given by relative accuracy, cover ratio, diversity and the number of acquired solutions. In the application to flowshop scheduling, heuristic decoding methods are proposed for each objective. Simulation result shows that the proposed method is effective to obtain diverse solutions especially in the problem with a large Pareto set, which results either from a statistical character of a given problem, or from a largeness of a problem scale, or from a high dimensionality of a solution space.
  • 最上 義夫, 馬場 則夫, 塩尻 晃, 田上 隆徳
    2002 年 38 巻 1 号 p. 97-103
    発行日: 2002/01/31
    公開日: 2009/03/27
    ジャーナル フリー
    In this paper, for variable hierarchical structure learning automata with S-model stationary random environment at each level, a new definition of optimal path is proposed based on the arithmetic mean of average rewards, and an LR-I type learning algorithm is constructed. The learning propertiy of our algorithm is considered theoretically, and it is proved that the probability of finding the optimal path can approach 1 as mach as possible by using our algorithm. In numerical simulations, the usefulness of our algorithm is shown.
  • 佐藤 洋平, 久村 富持
    2002 年 38 巻 1 号 p. 104-110
    発行日: 2002/01/31
    公開日: 2009/03/27
    ジャーナル フリー
    Switched arrival/server systems (SAS/SSS) are typical examples of hybrid systems which involve both continuous dynamics as well as discrete phenomena. It is known that SSS exhibits periodic behavior in asymptotic way while SAS may exhibit chaos. This paper treats SSS with three buffers where each buffer has capacity limitation. Focussing on periodic solutions with three switchings (P3), we analysed the generated periodic modes and their stabilities in detail. Finally, a control strategy for the periodic solutions P3 is proposed assuming that the capacity limits of buffers are available for control.
  • 青島 伸一, 吉澤 信幸, 藪田 哲郎, 羽成 賢一
    2002 年 38 巻 1 号 p. 111-113
    発行日: 2002/01/31
    公開日: 2009/03/27
    ジャーナル フリー
    We proposed the object discrimination system using a neural network with inputs for distance and sensitivity information of an ultrasonic sensor. We confirmed the validity of the system by the experiment using different grain size objects.
  • 鈴木 良昭, 鈴木 隆, 板宮 敬悦
    2002 年 38 巻 1 号 p. 114-116
    発行日: 2002/01/31
    公開日: 2009/03/27
    ジャーナル フリー
    This paper points out a problem that the adaptive gain matrix of the conventional constant-trace algorithm may lose its positive definiteness when the signal in adaptive control system is not persistently exciting and proposes a remedy for this problem.
feedback
Top