We envision highly flexible environments of
heterogeneous wireless networks of devices with different capabilities that
have access to multiple channels or network interfaces and roam over different
networks. In general, such environments are extremely complex and the
interaction of different layers and technologies creates many situations that
cannot be foreseen during the design and testing stages of technology
development. This is especially true for wireless networks, which are used for
many different purposes, and which are based on a shared medium that is
inherently more vulnerable than its wired counterpart. It is therefore critical
to perform comprehensive empirical studies in a wide range of production
environments to uncover deficiencies and identify possible optimizations and
extensions. The availability of high-quality measurement and modeling studies
would make it possible to develop wireless networks that are more robust,
easier to manage and scale, and able to utilize scarce resources more efficiently.
Our ultimate technological goal is to develop intelligent and robust wireless
network, which can be defined as network of devices that in a self-organizing
manner, monitor the environment, analyze its performance, and adapt based on
their resources and what was learnt to increase their quality of service. The
two essential components of this technology are (a) efficient, accurate and
scalable measurement and analysis techniques, and (b) adaptation mechanisms
that can optimize the performance and robustness of the entire system under
real workloads. The underlying philosophy in our research efforts is to tightly
integrate these two components by performing extensive real-life wireless
measurements, analysis, and modeling, and designing well-founded adaptation mechanisms. We will demonstrate the main
principles of our approach in the design of capacity planning, load balancing,
device adaptation, and location-sensing algorithms.
 Our research work uses real measurement data collected from large-scale
wireless networks. A significant part of the measured data is available
to the research community via the joint UNC/FORTH
repository of traces and models for wireless networks.
Mobile
peer to peer computing
To enhance the data availability, the group proposed a novel mechanism
that enables wireless devices to share resources. The focus is on three
facets of cooperation, namely information sharing, network connection
sharing and message forwarding. Peers communicate via a wireless LAN and
may have (intermittent) connection to the Internet e.g., via a Bluetooth
, modem or 802.11 AP. In the information sharing facet, peers query,
discover and disseminate information. When the network connection
sharing is enabled, the system allows a host to act as an
application-based gateway and share its connection to the Internet. The
group designed, prototyped and evaluated an architecture and set of
protocols, that enables this resource sharing in a self-organizing
fashion without the need of an infrastructure.
Collaborative Location Sensing (CLS)
Positioning is a
critical component of the mobile and pervasive, computing. The Mobile
Computing Activity at FORTH designed and evaluated the Cooperative
Location-Sensing (CLS) system that adaptively positions wireless-enabled
devices using the existing communication infrastructure (WiFi access
points) without the need of specialized hardware or training. CLS
employs the peer-to-peer paradigm enabling hosts to cooperate and share
positioning information. It also allows the easy incorporation of
external information (e.g., maps and spatial information, mobility
patterns) to improve its accuracy.
Research Statement [PDF]
|
