計測自動制御学会論文集
Online ISSN : 1883-8189
Print ISSN : 0453-4654
ISSN-L : 0453-4654
32 巻, 7 号
選択された号の論文の20件中1~20を表示しています
  • 原 進, 吉田 和夫
    1996 年 32 巻 7 号 p. 985-992
    発行日: 1996/07/31
    公開日: 2009/03/27
    ジャーナル フリー
    In the positioning control of flexible structures, it is important to develop a nonstationary robust control method for nonstationary uncertainty perturbations. At the same time, it is important to take account of influences of such unstructured uncertainties as ignored high-order vibration modes in order to suppress the spillover instability. This paper presents a nonstationary robust control method for flexible structures taking account of the structured and unstructured uncertainties influences by using a frequency-shaped time-varying criterion function. In order to avoid influences of unstructured uncertainties, in this method, a low-pass property is given to the controller from a frequency-shaped part of the criterion function. Moreover, an integrated design of both the H2 control for little influences of structured uncertainties and the H control for large influences is carried out by applying a time-varying part of the criterion function.
    In this paper, as an example, a positioning control problem of flexible inverted pendulum possessing infinity vibration modes is adopted. In this problem, the influences of structured uncertainties of Coulomb friction and parameter variations are generally severer in the period after positioning than in the period during positioning. In comparison with the former methods proposed by the authors, the usefulness of the present nonstationary robust control was verified by numerical calculations and experiments.
  • 藤井 裕矩, 松田 浩一, 中島 和成, 畠中 洋史
    1996 年 32 巻 7 号 p. 993-1000
    発行日: 1996/07/31
    公開日: 2009/03/27
    ジャーナル フリー
    Slew maneuver of a flexible space structure is studied in this paper to control both orientation and flextural vibration with using one control input only. The control method uses the concept of wave-absorbing control which is to absorb traveling waves in flexible structure at actuator positions. The wave-absorbing control is applied to suppress the excited vibration of flexible structure during slew maneuver and some significant requirements and limitations for the controller design are presented through analysis of the system using the traveling wave approach. Numerical simulation is used to analyze the performance in the frequency and time domains and satisfactory vibration suppression is shown in the low frequency band. Hardware experiment is also conducted to verify the validity of the numerical simulation and good agreement is shown in its performance of the present control algorithm.
  • 渡辺 亨, 吉田 和夫
    1996 年 32 巻 7 号 p. 1001-1010
    発行日: 1996/07/31
    公開日: 2009/03/27
    ジャーナル フリー
    In this paper, we present a design method of robust H/μ controller for flexible structures taking account of parameter variations and model errors in uncontrolled modes simultaneously. Parameter variations are modeled as structured uncertainties, and uncontrolled modes are replaced with high-pass filters which envelopes the dynamics of these modes. The uncertainties are implemented to the generalized plant as the weighting functions for state variables to reduce structural vibrations, and the filters are for control inputs to suppress feedback gain for uncontrolled modes, respectively. To investigate the effectiveness of the method, it is applied to the vibration control problem of a tower-like flexible structure which possesses transverse and torsional vibration modes. Computer simulations are carried out and it is shown that the H/μ controller synthesized according to the method achieves good robust stability and robust performance, while ordinaly LQG controller has poor robustness. It is also shown that the representation of structured uncertainties are optimized by multiplying D-scaling matrix through μ-synthesis. Control experiments are also performed and the robustness of the synthesized controller to the model errors of uncontrolled modes are confirmed. According to these results, the validity of the presented design method is certificated.
  • 松野 文俊, 笠井 正三郎, 田中 美和子, 若城 邦子
    1996 年 32 巻 7 号 p. 1011-1019
    発行日: 1996/07/31
    公開日: 2009/03/27
    ジャーナル フリー
    In this paper we consider robust force control of a one-link flexible arm. Since the tip of the flexible arm contacts with a given constraint surface, a constraint condition should be satisfied. By using Lagrange multiplier method and Hamilton's principle, we derive dynamic equations of the joint angle (an ordinary differential equation), the vibration of the flexible arm (a partial differential equation), and the constraint force. The boundary condition of the derived distributed parameter system is nonhomogeneous. We introduce a new variable and derive a homogeneous boundary condition. By solving an eigenvalue problem, eigenvalues and corresponding eigenfunctions are derived. On the basis of the eigenfunction expansion we derive a finite dimensional modal model. To compensate the spillover instability an optimal controller with low-pass property and a robust H controller are constructed. The stability of the direct force feedback and the compliance control considering uncertainty of the arm flexibility is analyzed. To demonstrate the validity of the derived model and the effectiveness of the proposed controllers, a set of experiments has been carried out. Experimental results confirm that the robust controllers perform well.
  • 土井 智晴, 大須賀 公一, 小野 敏郎, 川谷 亮治
    1996 年 32 巻 7 号 p. 1020-1026
    発行日: 1996/07/31
    公開日: 2009/03/27
    ジャーナル フリー
    To decrease a vibration with the robustness is necessary to control a flexible structure. In this paper, we discuss a robust vibration control for a time varying two-link-beam model regarded as a simple model of a flexible structure. We provide a design method of a robust vibration controller for the system. The controller is designed with, so-called, the design method of a constant-matrix-scaled quadratic stabilization. By this method, the H-infinity norm of the target transfer function is decreased, the gain of the designed controller becomes high and the controller becomes to have a derivative effect. Therefore, we can say that the resultant control system becomes a robust vibration control system for time varying systems. Finally, with some simulations, we show that the design method has effectiveness.
  • 山本 茂, 上田 隆弘, 木村 英紀
    1996 年 32 巻 7 号 p. 1027-1034
    発行日: 1996/07/31
    公開日: 2009/03/27
    ジャーナル フリー
    A design method to the benchmark problem provided by the robust control committee in SICE is considered. The model treated in the problem is a coupled three-inertia system which reflects the dynamics of the mechanical vibrations. The problem requires to be satisfied the robust performance (both the time-domain and the frequency-domain specifications). The feedback controller is designed based on the quadratic stabilization method which can deal with the robust performance problem. The feedforward controller is used to improve the time-domain specification simultaneously. Although the closed-loop system shows the complementary sensitive aspect, the other all design specifications can be achieved.
  • 西村 秀和, 大貫 修, 野波 健蔵, 戸谷 隆美
    1996 年 32 巻 7 号 p. 1035-1042
    発行日: 1996/07/31
    公開日: 2009/03/27
    ジャーナル フリー
    In this paper, for stabilizing and positioning control of a cart and inverted flexible pendulum system we formulate H servo compensator where displacement signals are used as the controlled variables and the cart velocity is used as feedback loop signal into the compensator. The feedback loop of the cart velocity which is excepted from the controlled variables is added through the loop gain to the feedback loop of the controlled variables in the design of the H servo compensator. By numerical calculations and experiments, we investigate the settling characteristics and the robust stability against a disturbance to the pendulum with two kinds of controllers, the output feedback controller by the displacement feedback and the output feedback controller adding the cart velocity feedback loop to the compensator. It is shown that although the displacement-output feedback controller without the cart velocity feedback loop has a big overshoot in the step response and the positioning of the cart is difficult, by means of the displacement-output feedback controller with a cart velocity feedback loop the overshoot can be reduced and stable positioning control of the cart is possible. Also it is verified that the servo compensator is superior to the regulator-based compensator in the settling characteristics.
  • 水野 毅, 石井 敏宏, 荒木 獻次
    1996 年 32 巻 7 号 p. 1043-1050
    発行日: 1996/07/31
    公開日: 2009/03/27
    ジャーナル フリー
    A self-sensing magnetic suspension system was studied both theoretically and experimentally. It used a switching power amplifier whose switching times were controlled by a current comparator with some hysteresis. An analysis and several experiments on the amplifier showed that its switching frequency changed linearly with the gap between the electromagnet and the suspended object. They demonstrated that switching signal could be treated as the output of a displacement sensor of frequency type. Phase-locked loop and frequency-feedback control were applied to stabilize the magnetic suspension system. In experiments, frequency-to-voltage (F/V) converters were used to realize the frequency-feedback control. Feeding back the F/V converted switching signal lead to stable suspension. When a phase-locked loop (PLL) control was activated in addition to the frequency-feedback control, the suspension system was pulled in lock. The accuracy of floating was improved by the PLL.
  • 北山 正文, 木村 文
    1996 年 32 巻 7 号 p. 1051-1056
    発行日: 1996/07/31
    公開日: 2009/03/27
    ジャーナル フリー
    The authors have proposed a three-dimensional measuring method used on particles of pulverized coal by the image processing system with two microscopes. The previous method was developed to measure the shape of pulverized coal by reproducing the horizontal cross section of pulverized coal at each height with the envelopment line based on projection pictures. However, the depressions of pulverized coal could not be measured accurately with this method.
    In this paper we propose a measuring theory which measures, using a laser spot, the three-dimensional shapes including the irregular surface conditions of pulverized coal. The measurement errors are estimated by simulation. The measuring results of the volume of pulverized coal by the present method and also particle shape coefficient φ are shown. Lastly, we compare this result with that of the three-dimensional measuring method used on particles of pulverized coal that was recently proposed by the authors.
  • 高橋 隆行, 大沢 博文, 鈴木 明宏, 猪岡 光
    1996 年 32 巻 7 号 p. 1057-1064
    発行日: 1996/07/31
    公開日: 2009/03/27
    ジャーナル フリー
    Based on the variation of the acceleration waveforms at the human waist, the authors have proposed an algorithm for identifying the human gait in their previously published papers. In this article, we discuss two points to improve the algorithm's performance: (1) Instead of using the zero-crossing point, the negative peak of vertical acceleration at the right side of waist is used to detect each step. This change makes the algorithm more robust to the variation of the cadence periods. (2) To compose the Two-Step waveform, two consecutive sets of the one-step-waveform are averaged. Using this method, ordering problem of the left/right foot data in composing Two-Step waveform is avoided.
    During the experiment, about 90% of gait have been identified by the improved algorithm successfully. We have also made a trial system which includes an interface for recording the identified gait into the Holter EGG.
  • 相原 伸一
    1996 年 32 巻 7 号 p. 1065-1073
    発行日: 1996/07/31
    公開日: 2009/03/27
    ジャーナル フリー
    We consider a nonlinear plant which is modeled by an Ito stochastic differential eqnuation. Corresponding to the plant, the nonlinear observation mechanism is aslo set where the observation noise is modeled by a finitely additive white noise.
    The main purpose of the present paper is to formulate the maximum a posteriori probability state estimate. By using the Onsager-Machlup functional, the a posteriori probability is derived. The basic equation for the MAP state estimate is derived with the aid of a dynamic programming approach.
    The numerical procedure for realizing the recursive filtering is also proposed.
  • 田中 幹也, 清水 顯
    1996 年 32 巻 7 号 p. 1074-1079
    発行日: 1996/07/31
    公開日: 2009/03/27
    ジャーナル フリー
    This paper presents a design scheme of robust model reference adaptive control systems (MRACS) to a linear time-varying plant. In this design scheme, the conventional direct method of MRACS is combined with a variable structure (VS) method. The influence of time-varying parameters of the plant is regarded as the disturbances to the time-invariant nominal model of the plant and is eliminated by using VS method. The effectiveness of this design scheme is demonstrated by the computer-simulations.
  • 長堂 勤, 示村 悦二郎, 石田 力, 内田 健康
    1996 年 32 巻 7 号 p. 1080-1086
    発行日: 1996/07/31
    公開日: 2009/03/27
    ジャーナル フリー
    In this paper, we propose a new controller reduction method which considers maintaining properties of the closed-loop system. Modern controller design techniques such as H-control, μ-synthesis frequently lead to high order controllers. Due to the practical reason, there is need to reduce thee order of these controllers. It is important to consider preserving the property of original controllers in the process of reducing the order. In the conventional method, an approximated criterion is used so that simple frequency weighted model reduction techniques may be applied. In our scheme, the structured singular value is introduced to derive a new criterion which has no approximation. A numerical example is used to confirm the effectiveness of the proposed method.
  • 中西 弘一
    1996 年 32 巻 7 号 p. 1087-1096
    発行日: 1996/07/31
    公開日: 2009/03/27
    ジャーナル フリー
    Though many calculating methods on steady state models of blast furnace were reported, a particular method always was applied in there. In order to explain the reason mentioned above, Iwao proposed “Basic Model of Simplified Blast Furnace” and applied two Goodness Criteria (rmsN, rmsD) to that. Furthermore, in order to analytically derive two Goodness Criteria, he showed that reaction products and reaction rate should be linearized by process data (reaction raw materials and reaction product analysis data).
    This paper proposed an approximate model of real blast furnace. This model was made up by reconsidering reaction raw materials and chemical reaction in the “Simplified Blast Furnace”: so that it complemented the following materials and chemical reaction.
    1) Injected fuel, water content, oxygen enrichment, SiO2, and CaO were complemented to the reaction raw materials.
    2) Ignition loss, reducing process of hydrogen, reduction of SiO2 and MnO, decomposition of coke and water content, and combustion of injected fuel were complemented to the chemical reaction.
    According to this model, the equation of relation between reaction raw materials and reaction products (I/O-BE: Input/Output-Basic Equation) could be derived. Furthermore, it was necessary that the calculating methods of reaction products should be chosen not material balance of C and O but all reaction product analysis data.
    Finally, it was shown that the calculating methods of reaction products could be linearized with quite good accuracy by process data.
  • 細江 繁幸, 赤坂 則之, 単 戈
    1996 年 32 巻 7 号 p. 1097-1106
    発行日: 1996/07/31
    公開日: 2009/03/27
    ジャーナル フリー
    The quality of wide film manufactured in industries is determined by the flatness of thickness along the width of the film. To control the film thickness in the inflated film process, uniform distribution of melted polymer flow velocity along the annular die gap is required and to realize this uniformity, many regulating heaters are used. The system is highly interracting since the heat generated by one heater makes changes of the film thicknesses at the adjacent heaters' positions due to the heat conduction in the die and usually more than several tens number of heaters are used. This makes it difficult to design the film control system.
    In this paper, a new design method is proposed for constructing inflated film thickness control systems. Here, the controller is assumed to be a PI type decentralized one, whose gain is determined by minimizing quadratic cost. The design is performed based on the following observed fact:
    By using the fact that the A-matrix of the closed-loop system comes to be a block symmetric circular matrix, it can be transformed into a block diagonal matrix by using similarity transformation which is completely determined only by the system size, not depending on the numeric characteristics of the process. This fact enables to decompose the large size matrix calculation into the small size one, and makes the stability test and the computation of cost feasible even for systems with the large number of inputs and outputs.
  • 深尾 典久, 川村 貞夫
    1996 年 32 巻 7 号 p. 1107-1112
    発行日: 1996/07/31
    公開日: 2009/03/27
    ジャーナル フリー
    To realize speedy and accurate motions of robot manipulators, it is well known that feedforward inputs are important. One of the method to realize the feedforward inputs is “Learning Control”. In such learning schemes, an ideal input torque pattern is formed through several actual iterative operations, only using input torque and output motion patterns of the previous trial. Therefore, the learning schemes can realize a desired motion easily. However, if we need many different motions, it is difficult to make the input torque patterns in the learning process, because the time and memory space to obtain and store the ideal input patterns are necessary. To overcome such difficulties, we propose a time scale translation for input torque patterns. Since translation of time scale means change of speed patterns of robot motions with a same spatial trajectory without using learning process. In the proposed translation, an arbitrary speed pattern is constructed only using four input patterns to realize four desired motions. Through some experimental results, the effectiveness of the proposed translation is demonstrated.
  • Taizo HANAI, Akihiro KAKAMU, Hiroyuki HONDA, Takeshi FURUHASHI, Yoshik ...
    1996 年 32 巻 7 号 p. 1113-1120
    発行日: 1996/07/31
    公開日: 2009/03/27
    ジャーナル フリー
    A modeling of total evaluation process of sake(Ginjo) was studied using a fuzzy neural network(FNN). Total evaluation of 61 Ginjo samples was estimated from each data set of 7 sensory evaluations. The values of performance index, J, based on the errors between actual and estimated values in FNN model were used for the evaluation of model. In FNN model with all 7 input variables, J value was 0.025 and it was almost similar to that of NN model as reported previously. The FNN model with 3 variables (color, flavor base and aging) selected previously was also tested, and J value (0.023) was also almost similar to that of NN model (J=0.024). To optimize the input variables and the number of membership functions, Parameter Increasing Method (PIM) was applied to the part of premise in FNN model. The FNN model obtained was constructed with 2 membership functions for color, 3 for flavor top, 2 for flavor base and 3 for hard-soft, and J value of 0.013 was fairy small. From analysis of connection weight of the FNN obtained, the acquired rules were easily described in the form of IF-THEN rule. Extraordinary flavor was added as a new input variable to above 3 FNNs. J values in all models decreased furthermore and especially the FNN with 3 variables and extraordinary flavor was found to show the lowest J value (0.010). It was concluded that 4 variables (color, flavor top, aging, extraordinary flavor) were important in the total evaluation of Ginjo. The results suggest that the fuzzy modeling using a FNN is effective on the analysis of sensory evaluation process.
  • 大藪 多可志, 木村 春彦
    1996 年 32 巻 7 号 p. 1121-1128
    発行日: 1996/07/31
    公開日: 2009/03/27
    ジャーナル フリー
    Indoor air-pollution monitorning system was developed using a semiconductive tin oxide gas sensor. The concentration of carbon dioxide has been used as an indicator for a grade of the air-pollution in Japan. However, this method has drawbacks because there are many polluting gases besides carbon dioxide in an indoor environment.
    In this study, the grade of pollution was relatively represented as the sensor output. Seasonal characteristics of the data obtained in a room for three weeks were calculated from daily output patterns with a computer. Its data were processed using OPS 5 and Rete algorithm to match condition elements of the rules and working memory elements. With this technique, temporal change such as season of the of f-set level of the sensor can be fairly negligible.
  • 湊 淳
    1996 年 32 巻 7 号 p. 1129-1134
    発行日: 1996/07/31
    公開日: 2009/03/27
    ジャーナル フリー
    School teachers must change seats of classroom every term. The simulated annealing is a method which can find a solution whose energy is minimum. In this paper, the simulated annealing method is applied to the problem of determination of seats of classroom which satisfies confusing conditions. Programs were developed and evaluated by computer simulations. As a result, it was found that the program can generate plural solutions which satisfies all conditions and program can be executed by a personal computer with practical time. The coolong schedule was also discussed.
  • 石井 千春, 申 鉄龍, 田村 捷利
    1996 年 32 巻 7 号 p. 1135-1137
    発行日: 1996/07/31
    公開日: 2009/03/27
    ジャーナル フリー
    It is well known that nonlinear H control problem is solvable if there exists a positive definite function which satisfies a certain Hamilton-Jacobi inequality. However, the routine method to obtain an analytic solution of Hamilton-Jacobi inequality is still unestablished. In this paper, we propose a method to obtain a solution of the Hamilton-Jacobi inequality by applying 'Exact Linearization via Feedback' for a nonlinear system. Then the solution of the Hamilton-Jacobi inequality is obtained as a quadratic form of a coordinate transformation with positive definite solution of a certain Riccati inequality.
feedback
Top