Kazuyoshi Murata

Low TCO and high-speed network infrastructure with virtual technology

In our university, we have updated our network infrastructure from ATM (Asynchronous Transfer Mode) network system to 10-Giga Ethernet system in March, 2009. The new system has 2 major features. One is ring topology with 10GBASE-X Ethernet. This feature enables lower initial cost than star topology because an ordinary star topology requires very expensive core switch system which can contain many 10GBASE-X GBICs (this device is also very expensive). The second is Virtual Chassis technology of Juniper Networks. This feature allows up to 10 switches to be interconnected and managed as a single, logical device. Thus it enables lower management cost than a system with many individual switches. In this paper, we show merits and demerits of our system against ordinary star topology.

Predicting Evidence of Understanding by Monitoring User's Task Manipulation in Multimodal Conversations

The aim of this paper is to develop animated agents that can control multimodal instruction dialogues by monitoring users behaviors. First, this paper reports on our Wizard-of-Oz experiments, and then, using the collected corpus, proposes a probabilistic model of fine-grained timing dependencies among multimodal communication behaviors: speech, gestures, and mouse manipulations. A preliminary evaluation revealed that our model can predict a instructors grounding judgment and a listeners successful mouse manipulation quite accurately, suggesting that the model is useful in estimating the users understanding, and can be applied to determining the agents next action.

When should animated agents give additional instructions to users? - Monitoring user’s understanding in multimodal dialogues -

In multimodal communication, verbal and nonverbal behaviors such as gestures and manipulating objects in a workspace occur in parallel, and are coordinated in proper timing to each other. This paper focuses on the interaction between a beginner user using a video recorder application on PC and a multimodal animated help agent, and presents a probabilistic model of fine-grained timing dependencies among different behaviors of different modalities. First, we collect user-agent dialogues using a Wizard-of-Oz experimental setting, and then the collected verbal and nonverbal behavior data will be used to build a Bayesian network model, which can predict the likelihood of successful mouse clicks in near future, given evidence associated with the status of speech, agents gestures and users mouse actions. Finally, we attempt to determine proper timing when the agent should give additional instructions by estimating the likelihood of a mouse click occurrence.

A web browsing method on handheld touch screen devices for preventing from tapping unintended links

In recent years, it is common thing to browse Web pages with mobile devices, such as smart phones. However, users sometimes tap the wrong link when they scroll or zoom web pages because of the relatively small display area of mobile device and sensitivity of touch screen. In such case, it is necessary to stop of loading the page or back to the previous page after the page changed. It seems that above unintended operation might increase the total browsing time and users frustration. In this study, we aimed to prevent users from tapping unintended links for effectively web browsing with touch-screen mobile devices. The proposed method has two kind of operation mode. They are a tapping mode and non-tapping mode. With the tapping mode, users can tap the link and change the mode only. On the other hand, with the non-tapping mode, users can do swipe, pinch, and mode change operation but they cannot tap any links. Furthermore, mode change operation, we adopt the Bezel Swipe operation, is intuitive and efficient. The results of the experimental evaluation showed that the rate of tapping the unintended links with the proposed method was lower than that with conventional method. However, the task completion time with proposed method is longer than that with conventional method.

Moodle integration of an automated account enabling system and a user status collection system

In this paper, we present a Moodle integrated system that combines an automated account enabling subsystem (AAES) and a user logon status collection subsystem in a computer lab (CULS). The AAES lets each new user learn how to use our computer system and the terms of use. After that the user can enable various computer services such as on-demand printing, network outlet service, and remote access after passing our automated examination in Moodle. The CULS shows information on users in the computer lab through our Moodle system. We have also implemented the capability of setting privacy levels for each role. For example, in the teacher role of a course, a user can get information about course-registered users, while the ordinary student role users can only get the number of free PCs available. The Moodle system content is available in both Japanese and English.

High-speed network infrastructure between KIT's campuses for computer system redundancy

In our university, we have updated our network infrastructure from Cisco-based 1-gigabit Ethernet system to Brocade-based 10-Gigabit Ethernet system in December 2012. We use a Brocade VDX Switch for aggregation Switch. It enables not only IP networking but also provides Fibre Channel and/or Fibre Channel over IP/Ethernet. Moreover, it provides connectivity between the Matsugasaki and Saga campuses (about a 10km distance) with 10GBASE-ZR Ethernet via our regional data center. It enables a campus-wide IP Storage Area Network (SAN), allowing computer systems to be operated with distributed environments, so our computer system will have redundant features against disasters. In this paper, we will discuss the merits and drawbacks of our system.

Effect of Haptic Perception on Remote Human-Pet Interaction

Even when a pet owner is away from his/her pet, he/she often wants to feel closer to the pet. The purpose of this study is to provide a means for the pet owner to feel the presence of his/her pet even when away from it; this is achieved by exchange of haptic feedback. In this paper, we describe such a remote haptic interaction system that consists of haptic devices for remote haptic communication; we also utilize tablets for video chat. A pet owner and his/her pet can feel closer to one another via haptic responses generated by corresponding haptic device. Two experimental evaluations were conducted to compare interactions between a pet owner and his/her pet using our system with another interactions achieved only via video chat. Results showed that these remote haptic interactions increased the pet owner’s feeling of communicating with his/her pet. In general, pet owners reported feeling closer to their pet by using our system.

A Displaying Method of Food Photos to Know Child’s Dietary Life for Parents

There are many children who live far away from parents. The parents want to know their child’s dietary life because it is important for his/her healthy life. However, there are no ways for parents to know their child’ dietary life without troublesome in each other. We proposed the method which extracts the information related to child’s meals automatically and displays it to know the child’s dietary life easily for parents.

An efficient document browsing method with floating diagram window on mobile device

This paper proposes an efficient method to browse the document including diagrams on mobile devices equipped with touch screen. With the method, the user can pull the referenced diagram close to the referrer sentence and browse the pulled-in diagram and the document simultaneously. The pulled-in diagram is displayed in the floating window overlapping the document. Both the pulled-in diagram and the underlying document are able to be scrolled and zoomed in/out by the user independently of each other. From the experiment, it is found that the proposed method is efficient for operation on mobile devices and the method is viewed very positively by the participants.

Empirical Evaluation of Throwing Method to Move Object for Long Distance in 3D Information Space on Mobile Device

In our previous work, a throwing method to move an object for long distance in 3D information space on a mobile device was proposed. With the method, as we throw the object to move it far away in the real world, we can throw the virtual object in 3D information space. This simple throwing method was improved by adding following three functions. They were real time adjusting the direction of moving object, moving viewpoint to follow the thrown object, and initializing viewpoint after the movement. The purpose of this paper is to examine the performance of the improved throwing method to move the object for long distance in 3D information space. From the experiment, it is found that the improved throwing method is efficient to move the object for long distance in 3D information space on mobile devices.