巻号一覧
16 巻, 2 号
選択された号の論文の5件中1～5を表示しています
- |<
- <
- 1
- >
- >|
-
隠れマルコフモデルと他の確率モデルの比較岡本 剛, 渡邊 裕司, 石田 好輝2003 年 16 巻 2 号 p. 61-69
発行日: 2003/02/15
公開日: 2011/10/13
ジャーナル フリーThe performance analysis is carried out for a masquerader detection system by a hidden Markov model (HMM). A Markov model and a probability distribution have been applied to many intrusion detection systems, however it is not clear to what extent the HMM improves the performance in detecting a masquerader, compared with other two probabilistic models. Statistics of these three models are compared and the advantage of the HMM for a masquerader detection system is clarified. Some implications useful in extending the other two probabilistic models to an HMM are also provided.抄録全体を表示PDF形式でダウンロード (1485K) -
浅田 武史, 中山 弘隆2003 年 16 巻 2 号 p. 70-76
発行日: 2003/02/15
公開日: 2011/10/13
ジャーナル フリーSupport Vector Machines (SVMs) are now thought as a powerful method for solving pattern recognition problems. SVMs are usually formulated as Quadratic Programming (QP). Using another distance function, SVMs can be formulated as Linear Programming (LP). In general, SVMs tend to make overlearning. In order to overcome this difficulty, the notion of soft margin is introduced. In this event, it is difficult to decide the weight for slack variables reflecting soft margin. In this paper, soft margin method is extended to Multi Objective Linear Programming (MOLP). It will be shown throughout several examples that SVMs reformulated as MOLP can give a good performance in pattern classification.抄録全体を表示PDF形式でダウンロード (932K) -
関森 大介, 臼井 智也, 升谷 保博, 宮崎 文夫2003 年 16 巻 2 号 p. 77-84
発行日: 2003/02/15
公開日: 2011/10/13
ジャーナル フリーIn this paper, we propose a method of self-localization for mobile robots. The method is based upon model-fittng of floor region provided by the omni-directional camera mounted on the robot. In our method, even if a part of the floor region is hidden by obstacles, this can be compensated for by computing the convex hull of a boundary point set of the floor region in the omni-directional image. The geometric features of the detected floor region and the linearized least square method considering the properties of omni-directional imaging are employed to fit the known floor shape. In addition, we developed a self-localization method that allowed for better estimates of values by integrating omni-directional vision and dead reckoning with the Kalman filter than by individual methods. Finally, we verify the effectiveness of these methods through several experiments with a real robot according to the rule of the RoboCup Small Size League.抄録全体を表示PDF形式でダウンロード (1299K) -
山田 功, 木下 航2003 年 16 巻 2 号 p. 85-93
発行日: 2003/02/15
公開日: 2011/10/13
ジャーナル フリーThis paper considers a design method of filtered inverse systems for time-invariant systems, which reproduce the input signal to the system approximately. The inverse system must be stable in practical applications. It is well known that the inverse system is unstable if the system has invariant zeros in the closed right half plane. Yamada and Watanabe presented a state space design method of stable filtered inverse systems for strictly proper systems with some right half plane invariant zeros. However, the series connected system with the system and filtered inverse system is not always diagonal rational function. The purpose of this paper is to give a state space design method of stable filtered inverse systems such that the series connected system with the system and stable filtered inverse system is a diagonal rational function. We construct preliminary the filtered inverse system by usual method and factorize the filtered inverse systems to several subsystems. The stable filtered inverse system is obtained using observer design for subsystems of the filtered inverse system. Numerical example is illustrated to demonstrait the effectiveness of the proposed method. Finally, an application of the stable filtered inverse systems is presented.抄録全体を表示PDF形式でダウンロード (1119K) -
浅井 徹2003 年 16 巻 2 号 p. 94-100
発行日: 2003/02/15
公開日: 2011/10/13
ジャーナル フリーThis paper deals with analysis of bumpy responses for systems with switching and uncertainty. Switching in control systems often causes bumpy responses. Since those responses are harmful, several methods have been proposed to attain bumpless transfer. However, the existing methods may not have enough robustness. Recently, another method has been proposed to analyze bumpy responses for systems with uncertainty, where the bumpy responses are analyzed through the worst-case L2 gain from the past disturbance to the future control output, with the expense of conservatism. This paper aims to improve the results by reducing the conservatism. The proposed analysis results are less conservative and given in terms of linear matrix inequalities. A numerical example shows that the proposed method actually gives a less conservative result than existing methods.抄録全体を表示PDF形式でダウンロード (932K)
- |<
- <
- 1
- >
- >|