Kato Laboratory at Graduate School of Information Sciences, Tohoku
University, is involved in research & development, education and
scholarly activities in the broad area of communication networks and
systems. Our research focuses on next-generation Internet and pattern
recognition applied technologies. The research areas of interest
include, but are not limited to wireless and mobile communications, Low
Earth Orbit satellite communication systems, network security,
development of ¡Èsmart¡É applications for next generation networks, and
character and document recognition using statistics and probability.
Following are brief introductions of the key research concentration
areas:
The field of wireless and mobile communications is one of the fastest
growing areas of the networking industry. Wireless communications
enable users to have access to their desired data beyond time and space
limitations, above all with higher degree of mobility. The research
topics of interests are listed hereunder:
Mobility management methods
Contents delivery techniques over heterogeneous
wireless networks
Mobile Ad Hoc Networks
Routing protocols for wireless networks
Development of schemes for guaranteeing Quality of
Service in wireless networks
- Low Earth Orbit Satellite Communication Systems -
Because of their extensive geographical reach, flexible and rapid
deployment features, and inherent multicast capabilities, satellite
network systems are seen as an attractive solution to realize the dream
of global personal communications. Given the recent advances and
ongoing improvements in satellite technologies, broadband satellite
based services are likely to open a promising and strong market for
service providers and operators. This research program includes
different issues in satellite networks, mainly those related to
frequent handover occurrences and unfairness among long and short RTT
connections. Our research in this topic includes:
Location-aware mobility management methods for
satellite networks
Contents delivery over satellite networks
TCP performance over satellite networks
Efficient and fair transmission protocols for satellite
networks
Design and development of an efficient integration of
satellites with terrestrial networks
- Network Security -
Recent events have illustrated that the underlying Internet
infrastructure is exposed to a high risk of malicious attacks. Without
adequate security, cyber terrorism may deprive large enterprises from
making efficient use of the Internet and defense organizations may fail
to guarantee the safety of their personnel in battlefield scenarios. To
secure the usage of the Internet, its vulnerability to these
illegitimate attacks should be investigated. This challenging task
underpins the research work of this area. Our research in this topic
includes:
Efficient and scalable tools for network traffic monitoring
Denial-of-Service Trace-back techniques
Ingress filtering techniques for attack preventions
- Next-Generation Network Applications -
With the recent developments of mobile computing and increasing
popularity of wireless communication, video conferencing or accessing
the Internet content from mobile devices (e.g. handset, pagers and PDA)
is no longer an imaginary scene. The following lists some topics we
investigate for the development of next generation ¡Èsmart¡É network
applications:
Wireless Application Protocols (WAP)
QoS and traffic classes in mobile networks
Human-oriented network interface applications
Context-aware and location-aware distributed services
¡ÈSmart¡É applications for high-speed and large-scale
mobile networks
- Pattern Recognition Technologies -
Developing intelligent computer systems with the ability of processing
information in a human-like manner is considered to be one of the most
challenging tasks in the computing area. Pattern recognition techniques
play a major role in realizing this vision. In Kato laboratory, we
search the usage of statistical and neural approaches in the
development of practical recognition systems. The following gives a
list of the topics under studies in this area:
Usage of multi-dimensional network information for the development
of efficient management applications
Usage of statistical information in the development of
highly accurate handwritten character recognition systems.