巻号一覧
48 巻, 11 号
選択された号の論文の14件中1～14を表示しています
- |<
- <
- 1
- >
- >|
<SSI2011特集―人間・社会・自然との調和を目指すシステム・情報技術―>
-
増田 士朗, 小林 孝一2012 年 48 巻 11 号 p. 687
発行日: 2012年
公開日: 2012/11/29
ジャーナル フリーPDF形式でダウンロード (149K)
論文
-
安部 原也, 伊藤 誠, 山村 智弘2012 年 48 巻 11 号 p. 688-697
発行日: 2012年
公開日: 2012/11/29
ジャーナル フリーIn order to clarify the effect of different time headways and lead car's decelerations on visual distraction and its impact on collision risk during car-following a driving simulator study was conducted. Moreover the experimental investigation considered system effectiveness of a forward collision warning system as for decreased collision risk. Results showed that compared to short time headway, long time headway induced long duration time of visual distraction during car-following. This effect was significant when a lead car decreased its speed mildly. Collision risk due to lead car's deceleration was low for long time headway during car-following compared to short time headway even thought visual distraction was induced. However there was a possibility that long time headway led to long duration time of visual distraction resulting in increased collision risk. In case of the alarm trigger logic used in this study, alarms effectiveness for decreasing collision risk was proven because early alarms were given for short time headway. It can be said that longer headway time does not always guarantee decreased collision risk.抄録全体を表示PDF形式でダウンロード (1545K) -
—仮想システムの生成範囲が再スケジューリングにもたらす影響—銭 毅, 藤井 信忠, 貝原 俊也, 藤井 進, 梅田 豊裕2012 年 48 巻 11 号 p. 698-704
発行日: 2012年
公開日: 2012/11/29
ジャーナル フリーIn this study, a new concept of agent-based Real-Virtual Fusion Manufacturing System (RVF-MS) is proposed, which aims to adaptively and effectively deal with both external and internal fluctuations by realizing a fusion between real production shop floor and virtual manufacturing system model. In this paper, for the social contract-based production scheduling in RVF-MS, a method of limiting partial components for generating virtual system is proposed at operational phase. The influence of the generated range of virtual systems on rescheduling results is analyzed by computational experiments on flexible flow shop problem.抄録全体を表示PDF形式でダウンロード (535K) -
山口 拓真, 稲垣 伸吉, 鈴木 達也2012 年 48 巻 11 号 p. 705-712
発行日: 2012年
公開日: 2012/11/29
ジャーナル フリーThis paper presents construction strategy of Bayesian Network (BN) structures in decentralized fault diagnosis of event-driven systems based on probabilistic inference. In this decentralized diagnosis method, a fault is identified using the BN and Timed Markov Model (TMM). The BN represents the causal relationship between the faults and the observed event sequences in subsystems, and the structure of the BN is essential since the computational complexity and the fault diagnosis performance depend on it. Therefore, this paper proposes construction strategy of the BN based on an importance indicator of arc, which expresses independence properties between faults and observations, in fault diagnosis of event-driven systems. Finally, the usefulness of the proposed strategy is verified through some experimental results of an automatic transfer line simulated on a PC.抄録全体を表示PDF形式でダウンロード (1142K) -
中田 雅也, 原田 智広, 佐藤 圭二, 松島 裕康, 高玉 圭樹2012 年 48 巻 11 号 p. 713-722
発行日: 2012年
公開日: 2012/11/29
ジャーナル フリーThis paper focuses on Identification-based XCS (IXCS) which introduces the identification mechanism into XCS (Accuracy-based Leraning Classifier System) and extends it to Predicted reward-based IXCS(PIXCS) to promote a generalization of classifiers(i.e., rules) in the binary and multi-classification problems with reducing the number of classifiers. Through the intensive simulation of 20-Multiplexer problem and 3×3 Concatenated multiplexer problem, this paper has revealed the following implications which cannot be achieved by the conventional LCS(i.e., XCSTS) and IXCS: (1) PIXCS can derive better performance than XCSTS and IXCS in the binary-classification problem, (2) PIXCS can generalize not only the classifiers faster than IXCS but also the classifiers which are robust in the noisy multi-classification problem with reducing the number of classifiers.抄録全体を表示PDF形式でダウンロード (1545K) -
大野 文夫, 高橋 大志2012 年 48 巻 11 号 p. 723-730
発行日: 2012年
公開日: 2012/11/29
ジャーナル フリーThis paper analyzes the corporate strategy on environmental regulations through Agent-Based Modeling. Once a country decides to implement environmental regulations on green policy, the target companies should comply with the regulations. However, the companies may take another strategy that does not always comply with the regulations, for extra costs are likely to be needed to comply with the regulations. In this analysis, we simulate the corporate activities for environment-conscious products (ECP), consumer's purchasing patterns and the direction of a country for penal regulations on environmental regulations through Agent-Based Modeling. As a result, on the assumption that consumers always purchase from the companies that have been less punished in two random selected companies, we found (1) the companies that always comply with the regulations are not likely to exist in the end of period without penal regulations, (2) the length of company's strategy has less influence on the company's strategy, (3) the change of consumer's purchasing patterns is important to effectiveness of the environmental regulations.抄録全体を表示PDF形式でダウンロード (1478K) -
平野 博大, 堀内 徹也, 平野 陽豊, 栗田 雄一, 鵜川 貞二, 中村 隆治, 佐伯 昇, 吉栖 正生, 河本 昌志, 辻 敏夫2012 年 48 巻 11 号 p. 731-739
発行日: 2012年
公開日: 2012/11/29
ジャーナル フリーThis paper proposes a novel technique to monitor peripheral vascular conditions by using biological signals, such as electrocardiogram, arterial pressure, and pulse oximetric plethysmogram. A second-order log-linearized model (called the log-linearized peripheral arterial viscoelastic model) is used to describe the nonlinear viscoelastic relationship between the blood pressure waveform and the transillumination plethysmographic waveform. The proposed index is able to estimate the changes of stiffness of peripheral arterial wall induced by sympathetic nervous activity, and the validity of the proposed method is then discussed by monitoring peripheral vascular conditions during arm position tests and during endoscopic thoracic sympathectomies (ETSs). As results of the arm position tests, the stiffness was arm position-independent (Up: 4.0 [%], Down: 5.5 [%]). Then, as results of the ETSs, the variation of the stiffness was significantly changed between before and during the ETS procedure (p < 0.01), and between during and after the ETS procedure (p < 0.01). The above experimental results clearly show that the proposed method can assess changes in the sympathetic nervous activity during ETSs.抄録全体を表示PDF形式でダウンロード (2014K) -
松本 裕樹, 武村 紀子, 中村 泰, 岩井 儀雄, 石黒 浩2012 年 48 巻 11 号 p. 740-744
発行日: 2012年
公開日: 2012/11/29
ジャーナル フリーIn this paper, we propose a control method for multiple lights to provide a comfortable living environment for all residents in it. It is not easy to configure the illuminance environment because of dynamic changes in the environment which are induced by disturbances such as obstacles of light and the external light. In this study, we employ the light occlusion rate parameter to estimate the illuminance situation and the voltage of light is controlled based on the estimated value. Furthermore, we employ a control rule based on the theory of biological fluctuation to make the estimation more precisely. Experimental results show that our method works well and the occlusion parameter can be estimated stably.抄録全体を表示PDF形式でダウンロード (1726K) -
服部 聖彦, 高玉 圭樹, 大谷 雅之, 松島 裕康, 佐藤 圭二, 市川 嘉裕2012 年 48 巻 11 号 p. 745-753
発行日: 2012年
公開日: 2012/11/29
ジャーナル フリーThis paper proposes a distributed and self-organized method which has the mechanisms of robustness sensor data collection and fault diagnosis for large-scale sensor networks. To cope with complex sensor networks which include (i) an ad-hoc multi-hop communication, (ii) data aggregation, and (iii) high error data rate caused by the a huge number of nodes, the proposed method collects the information of nodes and diagnoses its nodes by introducing the new nodes called token nodes and the new type of broadcast called a limited virtual broadcast. The token node can create packets which can only be transmitted in the limited multi-hop networks as a limited virtual broadcast, which can prevent a broadcast storm. To clarify the effectiveness of our method, we conduct the simulations on two-hops network and compare the results of our method with those of the conventional method called Adaptive DSD. The intensive simulation results have revealed that our proposed method works well in the large-scale sensor network and shows better performance than Adaptive DSD method from the viewpoint of the communication traffic and rate of information sharing.抄録全体を表示PDF形式でダウンロード (1562K) -
平岡 敏洋, 西川 聖明, 川上 浩司, 塩瀬 隆之2012 年 48 巻 11 号 p. 754-763
発行日: 2012年
公開日: 2012/11/29
ジャーナル フリーEco-driving is regarded as a simple and effective countermeasure to reduce fuel consumption. However, the execution and continuity of the eco-driving depend extensively on a driver's motivation because it requires more complex maneuvers than the conventional driving. Therefore, we propose a novel eco-driving support system based on the psychological theories with respect to motivation, that intends to encourage the driver to perform spontaneous fuel-efficient driving behavior. Furthermore, driving simulator experiments were conducted to verify the effectiveness of the proposed system.抄録全体を表示PDF形式でダウンロード (1511K) -
市川 嘉裕, 高玉 圭樹2012 年 48 巻 11 号 p. 764-772
発行日: 2012年
公開日: 2012/11/29
ジャーナル フリーThis paper focuses on the conflict which occurs even in the cooperated task in multi-agent reinforcement learning domain and explores its avoidance methods from the viewpoint of the learning progress of agents. For this purpose, we propose the conflict avoidance method in multi-agent Q-learning which enables agents to learn according to the learning progress of other agents, and investigate its effectiveness in the lane crossing road problem where agents aim at going to their own directions but they cannot cross each other due to the narrow lane. The intensive simulations on the problem has the following implications: (1) the temporary decrease of the discount factor of the Q-learning agent with a high learning progress (i.e., the agents who learns well) contributes to avoiding conflict more successfully than the increase of the random degree of agent behaviors; (2) the increase of the discount factor of the Q-learning agent with a low learning progress (i.e., the agents who has not yet learned well) contributes to avoiding the conflict more successfully; and (3) a balance of learning progress among agents is critical to solve the conflict situation.抄録全体を表示PDF形式でダウンロード (1336K) -
山下 泰央, 上瀧 弘晃, 高橋 大志2012 年 48 巻 11 号 p. 773-780
発行日: 2012年
公開日: 2012/11/29
ジャーナル フリーThis research analyzed the relationship between headline news and bond market in Japan through text mining technology. As a result of intensive analyses, we found the significant relationship between interest rate changes and the negative (or positive) news classified by keywords. At the same time, we also found that a measure of constructing keyword list has a significant impact on performance of classification.抄録全体を表示PDF形式でダウンロード (565K) -
鈴木 宏典2012 年 48 巻 11 号 p. 781-789
発行日: 2012年
公開日: 2012/11/29
ジャーナル フリーA dynamic feedback system is developed to estimate velocity and headway distance in a longitudinal three-vehicle platoon. The estimation system is modeled by extended Kalman filter (EKF) as well as neural Kalman filter (NKF) that estimate the velocity and headway distance by measuring acceleration rate of some selected vehicles in the platoon. State equations of the EKF are analytically defined by discrete conservation equation of vehicle speed and spacing, whereas the measurement equation is based on a conventional car-following model. The NKF, however, defined both equations by artificial neural network models (ANNs) which enables both equations to be defined without using any analytical equations. Numerical analysis showed that the NKF reduces the estimation errors in most cases compared to EKF because of the high capability of ANN models for describing non linear phenomena. However, less statistical difference was observed between NKF and EKF due to the lack of data sets or measurement variables.抄録全体を表示PDF形式でダウンロード (1586K) -
飯間 等, 黒江 康明2012 年 48 巻 11 号 p. 790-798
発行日: 2012年
公開日: 2012/11/29
ジャーナル フリーWe recently proposed swarm reinforcement learning methods in which multiple sets of an agent and an environment are prepared and the agents learn not only by individually performing a usual reinforcement learning method but also by exchanging information among them. The methods have been applied to problems with discrete state-action space as the first stage of the research. In the real world, however, there are many problems which are formulated as ones with continuous state-action space. This paper proposes swarm reinforcement learning methods for acquiring optimal policies rapidly for problems with continuous state-action space. The information exchange methods which we proposed for the discrete problems can not be directly applied to continuous problems in which the state space is higher-dimensional and the state value function has stronger nonlinearity. We propose new information exchange methods which can be applied to such continuous problems. The proposed swarm reinforcement learning methods are applied to a biped robot control problem, and their performance is examined through numerical experiments.抄録全体を表示PDF形式でダウンロード (568K)
- |<
- <
- 1
- >
- >|