巻号一覧
39 巻, 2 号
選択された号の論文の15件中1～15を表示しています
- |<
- <
- 1
- >
- >|
-
平井 宏明, 升谷 保博, 宮崎 文夫2003 年 39 巻 2 号 p. 89-96
発行日: 2003/02/28
公開日: 2009/03/27
ジャーナル フリーIn order for a robotic system autonomously to capture an unknown rigid body rotating freely in zero gravity, it is necessary for it to detect the object's motion. This paper presents a method of estimating object motion based on position information concerning a point located on the object's surface. We assume that the object's motion is described in term of the combination of the rotations with constant angular velocities. When position sensors, etc., can obtain three dimensional position information about a point on the object, we can estimate parameters describing its motion. In other words, using this method we can calculate each frequency and direction of the axes of rotation by applying the Fourier transform to a time sequence of the point's three dimensional position. This is also useful in case temporal sequence of the observed point has interruption owing to occlusion, etc..抄録全体を表示PDF形式でダウンロード (2016K) -
喜安 千弥, 藤村 貞夫2003 年 39 巻 2 号 p. 97-103
発行日: 2003/02/28
公開日: 2009/03/27
ジャーナル フリーSpectral unmixing is a method by which to estimate the proportion of each component in a pixel using multispectral data. In conventional analysis of remotely sensed images, each pixel is classified into a single object category. However, the actual land surface corresponding to a pixel does not necessarily consist of only one category of objects. Therefore, estimating the proportion of components that exist in a pixel is often useful. The most commonly used method of spectral unmixing assumes that the component spectra are determined from training data. However, available training data do not always correctly represent the spectral characteristics of the categories within the objective area. In such cases, large errors may appear in the results of unmixing.
We propose herein the adaptive spectral unmixing method, which estimates suitable component spectra from the actual observed data and thus requires no training data. By adaptively estimating the component spectra from the set of observed data in the objective area, we can correctly estimate the proportion of components even if the spectral characteristics change with the location of objective area. In the proposed method, the spectral reflectance of pixels is expressed by vectors in multi-dimensional space, which can be written as linear combinations of component spectra weighted according to component proportion. We determine the component spectra by finding the minimum volume of simplex containing all of the reflectance vectors, where the vertexes of the simplex correspond to the component spectra.
We estimated the degree of errors by numerical simulation and compared the performance of the proposed adaptive method and that of the conventional method. We confirmed that the proposed method of adaptive unmixing provides better results than the conventional method when the spectral characteristics change with the location of the objective area.抄録全体を表示PDF形式でダウンロード (1368K) -
瀬戸 学雄, 信太 克規2003 年 39 巻 2 号 p. 104-109
発行日: 2003/02/28
公開日: 2009/03/27
ジャーナル フリーAuthors introduce here the plane type capacitance sensor for the human body detection technology used for realization of nursing care equipment. Since this sensor has characteristic form, it is used installing in equipment. In this sensor, it is necessary to distinguish the big difference of permittivity between human body and general objects on a seat, even when cushion material exists between objects and sensor in practical use. The useful results of simulations for the sensor properties with dummy electrodes are presented. The capacitance values in a 2-dimensional model were calculated with the FEM analysis. It becomes possible to acquire high capacitance by setting up dummy electrodes in the inner side and outside of measurement electrodes in the plane type capacitance sensor, and an object can be distinguished according to the difference of the capacitance by the difference in the relative permittivity. Furthermore, the sensor that we proposed showed the detection possibility of the size of an object.抄録全体を表示PDF形式でダウンロード (2436K) -
不破 勝彦, 成清 辰生, 神藤 久2003 年 39 巻 2 号 p. 110-119
発行日: 2003/02/28
公開日: 2009/03/27
ジャーナル フリーIn this paper, a design method of control systems which achieves disturbance rejection for the multiple-input systems using a dual observer is proposed. This method is extended our earlier work to more general case and has the following advantages.
(1) An unobservable subspace of arbitrary dimension in the state space could be made by dual observer.
(2) In strictly multiple-input systems, poles of the unobservable subspace could be assigned arbitrarily.
Therefore, according to the construction of the unobservable subspace using this method, the disturbance rejecting systems could be constructed. In this regards, the validity of constructed control system is shown by the numerical simulations for 3-mass spring system.抄録全体を表示PDF形式でダウンロード (1962K) -
Manabu YAMADA, Yasuyuki FUNAHASHI2003 年 39 巻 2 号 p. 120-126
発行日: 2003/02/28
公開日: 2009/03/27
ジャーナル フリーThis paper presents an explicit parameterization of all deadbeat tracking controllers for multi-input multi-output (MIMO) control systems in one-degree of freedom compensation scheme. The advantages over the previous results are as follows: Firstly, both all deadbeat tracking controllers and all achievable deadbeat responses are characterized by completely free parameters in an explicit form. Moreover, the relationship between the number of the settling step and the free parameters is made clear. Secondly, the class of the reference inputs for which deadbeat tracking is achieved is generalized. Thirdly, the design method is obtained not only for the case that the plant matrix is square but also for the non-square case. Fourthly, the design algorithm consists of simple polynomial matrix equations and the reduction to Hermite form over polynomials, while the previous algorithms in one-degree of freedom compensation scheme involve some troublesome calculations, e.g., the reduction to Smith-McMillan form of rational functions.抄録全体を表示PDF形式でダウンロード (1620K) -
岩村 誠人, 河野 裕介, 尾崎 弘明, 毛利 彰2003 年 39 巻 2 号 p. 127-135
発行日: 2003/02/28
公開日: 2009/03/27
ジャーナル フリーThis paper presents a novel gradient-type iterative algorithm for the solution of optimal control problems with constraints on the control variables, the terminal state and the state variables. In this algorithm, the control variation for improvement is constructed in a hierarchical manner so that it decides the convergence rates of the terminal constraints, the state constraints and the cost functional separately, to improve the convergence performance. A clipping-off technique is used to deal with the input constraints. The proposed algorithm can be also applied to the problem with final time unspecified by introducing a time-scale factor. A numerical example is shown to illustrate the effectiveness of the proposed algorithm.抄録全体を表示PDF形式でダウンロード (1752K) -
古屋 栄彦, 小林 伸明, 鈴木 亮一, 中溝 高好2003 年 39 巻 2 号 p. 136-142
発行日: 2003/02/28
公開日: 2009/03/27
ジャーナル フリーIn this paper, a solvability condition of an exact model matching problem for a structurd descriptor system is studied by using a graph theory.
Several researche studies have been carried out disturbance decoupling conditions for a structurd system in a state space form. Normally, a state space form is obtained by a system transformation of a descriptor form; however, representations of mathematical structures do not have physical structures of a real system if there exists dependencies between elemant of matrices on the transformation procedure. Thus we discuss the structured descriptor system in here.
Firstly, we consider a solvability condition of the disturbance decoupling problem for the descriptor system, and formulate an exact model matching problem. Finally, we apply the result of structural disturbance decoupling problem to the exact model matching problem for a structured descriptor system, and show its solvability condition.抄録全体を表示PDF形式でダウンロード (1654K) -
Satoru TAKAHASHI, Bijoy K. GHOSH2003 年 39 巻 2 号 p. 143-149
発行日: 2003/02/28
公開日: 2009/03/27
ジャーナル フリーIn this paper, we consider the problem of estimating motion and shape parameters of a plane undergoing three different motion models using both a CCD camera and a laser range finder. The camera observes features on the moving plane perspectively or orthographically. The laser range finder observes a cross section of the plane as a line. We use this time-varying line information for the estimation problem of motion and shape parameters, and obtain the extent to which motion and shape parameters can be estimated.抄録全体を表示PDF形式でダウンロード (1698K) -
ハイブリッドシステムの運動計画八島 真人, 山口 秀谷2003 年 39 巻 2 号 p. 150-158
発行日: 2003/02/28
公開日: 2009/03/27
ジャーナル フリーThis paper presents the motion planning for enveloping grasp with switching contact modes. We should consider that the dynamics and kinematics change according to the contact modes. We systematize the properties of the manipulation system with switching contact modes based on its dynamics and kinematics, and discuss the mathematical conditions to be satisfied for its motion planning. The conditions give the restriction of the feasible contact mode and the number of contact points. Noticing that this manipulation system belongs to a class of a hybrid system, we propose a layered motion planning algorithm. We finally show the example of the motion planning for enveloping grasp.抄録全体を表示PDF形式でダウンロード (2377K) -
神藤 良康, 岩瀬 将美, 畠山 省四朗, 古田 勝久2003 年 39 巻 2 号 p. 159-167
発行日: 2003/02/28
公開日: 2009/03/27
ジャーナル フリーIn this paper, an identification algorithm for continuous-time systems is proposed. It provides the minimum state space realization of a plant directly from input-output data sampled at regular intervals. It needs the highorder derivatives of input-output signals, but to obtain them is known to be difficult in general. Therefore, a filter and its derivatives are introduced, and the output of the filters can be considered as the derivatives of intput-output. As the result, the precise derivatives of the input-output can be obtained as the output of the filters. If a low-pass filter is chosen as the filters, it can be expected that high-frequency noise is attenuated. Furthermore, the order of the needed derivatives is comparatively lower because this method is based on the Gopinath-Budin's realization algorithm for discrete-time systems. The accuracy of identification is also improved by introducing the input-output scaling and the instrumental variable. Comparing the proposed method with the existing works through numerical simulations, its effectiveness is verified.抄録全体を表示PDF形式でダウンロード (2294K) -
古月 敬之, 平澤 宏太郎2003 年 39 巻 2 号 p. 168-175
発行日: 2003/02/28
公開日: 2009/03/27
ジャーナル フリーNeural networks have attracted much interest in system identification and control communities because they can learn any nonlinear mapping. However, from a user's point of view, when neural networks are used as models for controller design, they do not have structures of easy use. This paper introduces a new neural network based prediction model for control of nonlinear systems. Distinctive features of the new model to conventional neural-network based ones are that it has not only meaningful interpretation on part of its parameters but also is linear for the input variables. The former feature makes parameter estimation easier and the latter allows us to derive a nonlinear controller directly from the identified prediction model. The modeling and the parameter estimation are described in detail. The usefulness of the new model is demonstrated by using numerical simulations.抄録全体を表示PDF形式でダウンロード (1703K) -
松村 嘉之, 大倉 和博, 上田 完次2003 年 39 巻 2 号 p. 176-184
発行日: 2003/02/28
公開日: 2009/03/27
ジャーナル フリーThis paper experimentally investigates the effect of recombination operators on multirecombinant Evolution Strategies ((μ/μ, λ)-ES). For ES, each individual has a pair of real-valued vectors, i.e., object parameters and strategy parameters. We adopt two popular recombination operators, namely Intermediate Recombination (IR) and Discrete Recombination (DR), and apply IR or DR to both object parameters and strategy parameters. Computer simulations are conducted using three types of (μ/μ, λ)-ES, which are Classical-ES (CES), Fast-ES (FES) and Robust-ES (RES). We show the advantages of IR or DR in CES, FES and RES. Especially, DR on a pair of an object parameters and a strategy parameters improves the performance of all ESs in our experi-ments. By means of principal components analysis of evolutionary dynamics, we discuss the characteristics of (μ/μ, λ)-ES.抄録全体を表示PDF形式でダウンロード (1966K) -
飯間 等, 三宮 信夫2003 年 39 巻 2 号 p. 185-191
発行日: 2003/02/28
公開日: 2009/03/27
ジャーナル フリーLocal search methods are effective for solving scheduling problems in the case where the computation time is limited. However, the performance of the local search methods depends on the definition of neighborhood, that is, the procedure to generate new solutions. In this paper, a local search method with new neighborhood structures is proposed for job shop scheduling problems of minimizing the makespan. The solution in the proposed method is represented by an operation sequence. In this representation a same schedule is generated generally from different operation sequences, and such redundancy reduces the search efficiency. The proposed neighborhood structures generate necessarily a schedule which is different from the incumbent one. Moreover, two neighborhood structures based on this idea are used by turns for escaping from a local optimum. The proposed method is applied to some benchmark problems. It is shown from the computational results that the proposed method outperforms other local search methods.抄録全体を表示PDF形式でダウンロード (1558K) -
凌 元錦, 幸田 武久, 鈴木 和彦, 島田 行恭2003 年 39 巻 2 号 p. 192-198
発行日: 2003/02/28
公開日: 2009/03/27
ジャーナル フリーThis paper studies the reliability and safety of the k-out-of-n: G system as well as its optimal structure. Considering intermittent failure and complete failure, when the expected loss (I) caused by fail-safe (FS) and faildangerous (FD) is used as the evaluation criteria of safety, the relationship between reliability and safety is expressed by an analytical formula. It is analyzed mathematically that the high reliability cannot always guarantee its safety. We discuss how the expected loss, the probability of FS and the probability of FD vary, depending on the reliability of sensor. Further, a simple formula is obtained for the optimal κ that can minimize I. The change of the optimal dynamic logic structure as time of the system elapses is argued. Illustrative examples are given.抄録全体を表示PDF形式でダウンロード (1525K) -
西川 郁子, 中澤 重人, 喜多 一2003 年 39 巻 2 号 p. 199-208
発行日: 2003/02/28
公開日: 2009/03/27
ジャーナル フリーA coupled system of phase oscillators is proposed for a signal control of an urban traffic network. An area-wide control is given by a network of autonomous systems, which exchange local information without any central monitoring system. Each signal is modeled by a simple phase oscillator with a periodic motion, and pair-wise interactions with the adjacent signals control its offset to attain a desired value. The interaction strength depends on a traffic flow density between the corresponding pair of adjacent signals. The proposed control method is applicable to a rather general traffic flow including a loop road, a one-way road, left or right turn traffic at an intersection and any number of roads at an intersection. The results of computer simulations on several traffic flow patterns and conditions in a system of 2×2 signals are shown both with a macro model described by a flow density and with a micro model considering the dynamics of an individual car. Statistics such as an average travel time show that the effective control is achieved adaptively by the proposed method compared with several simple offsets in most conditions.抄録全体を表示PDF形式でダウンロード (2399K)
- |<
- <
- 1
- >
- >|