Professor, Division of Spatial Information Engineering

Further information is available at http://www.mcl.iis.u-tokyo.ac.jp/

Context aware services

This research aims at constructing an architecture for flexible context aware services, where the intent of users are proactively recognized and their activities are properly and comfortably supported. Among them, physical location of users is the most important “context”. Thus, in this research we especially focus on the location dependent routing and network services. While the conventional routing of internet and PSTN both depend on the logical user address as IP address and telephone number, we are studying the routing architecture which depend on the physical location of the users and the intermediate nodes. By using it, we can deliver necessary information to the user of a specific physical location. We are also studying the enhancement of this architecture to the multicast case which is called geo-cast. By geo cast, we can deliver necessary information to the users of a specific geographical region, say within 100m from Tokyo station. Based on location dependent routing, various network services can be provided. For example, in an emergency, refuge alert can be promptly delivered to the proper geographical location, which might help the mitigation of disaster such as earthquake and fire.

Haptic collaboration and its applications

Communication via networks has been based on voice and image media for over 100 years. On the other hand, various haptic devices, by which the sense of touch and force are sensed and fed back to the users at the same time, is now available as commercial products. Thus haptic communication is now becoming realistic and applications using them are also emerging and expected as one of the promising killer applications. However, since haptic collaboration and communication have been studied mainly by researchers of robotics, the effect of networks connection haptic devices have been ignored. In other words, the networks have been treated as ideal. But, in reality, the quality of the haptic collaboration highly depends on the network impairments.
From this point of view, we are studying various issues related to keeping the subjective quality of haptic applications as high as possible under such network impairments as bandwidth limitation, latency, jitter and loss of information. The examples are coding of haptic media, media synchronization for the compensation of the delay and jitter, various method for error correction including dead-reckoning.

Study on high-speed networks

We have been studying the various topics on the operation of very high speed networks. As for the network elements, we are interested in very high speed node systems as a very lager scale ATM switch fabric and optical switch for WDM networks. From the viewpoint of network planning, we proposed the optimization method of the fiber network where the dominant traffic is broadband multicast/broadcast, the dynamic distribution algorithm of VOD servers and fast wavelength path assignment algorithm for WDM networks. We have been also studying the protocol design of layered transmission and layered multicast. Transparent media scaling by the combination of above protocols and layered coding of images have been also proposed.