Early research interests include the theory of concurrency, with an emphasis on the aspects that relate interleaving and causality, and the theory of parallel computation.
Brief description. Two major and lasting trends in
the networking landscape are the growing importance of
wireless networks and the increasing diversity of wireless
networking solutions and standards. These trends come hand
in hand with the rapidly decreasing cost and shrinking
physical size of radio interfaces. It is now technically and
economically feasible to put several radio
transmitters/receivers in a single wireless network
node. This creates an environment where the network
effectively has multiple physical layers. This is expected
to become ubiquitous in the future. While the technical
possibility of multiple physical layers is already quite
clear today, it is much less obvious how can this
opportunity be efficiently utilized to gain significant
improvement in the network performance. Or, from the
practical/economical point of view, the ultimate question
is: will the multiple physical layer (multi-radio) network
development lead to sufficient performance improvement that
justifies the investment?
In this research project the
investigators develop and analyze novel mathematical methods
that can can quantify the network performance gain that is
obtained via multiple physical layers. Specifically, the
investigators model the network topology with an
edge-labeled multigraph. This model offers surprisingly
richer opportunities than the traditional graph model. Using
this approach, the investigators study the following main
areas: (1) quantifying the multi-radio gain in the network
topology; (2) new algorithmic problems at the network layer;
(3) new issues in network reliability; and (4) modeling and
choosing routes in a mobile environment (5) experimental
validation of the results via a testbed built in the
project.
Brief description. The proposed research is aimed at using the technology of Integer Linear Programming (ILP) for exploring controlled mobility in wireless networks. We consider multi-hop wireless networks were a large number of nodes are statically placed and only some of them can move. By using ILP techniques we want to show that routes and schedules for the mobile nodes can be found that optimize crucial network performance metrics, such as the network lifetime and the end-to-end data packet latency. Contributions of the proposed research are multifold. We expect to advance the state of the art in wireless networks where some nodes are mobile. Furthermore, we will test and push the current capabilities in ILP modeling and solution technology. Among the expected outcome of the proposed research we will define complex models of realistic (in size, parameters, etc.) network scenarios; we will develop and test heuristics to make the ILP formulations more scalable; we will determine provable performance bounds on metrics of interests for wireless network, and we will compare heuristic solutions to these bounds for rigorous benchmarking and protocol design and optimization.
Brief description. This research focuses on exploring
solutions that will allow small, energy-constrained wireless
sensor nodes to self-localize, i.e., to compute their
coordinates with respect to an absolute or relative
positioning system. Typical wireless sensor networks (WSNs)
applications that require accurate localization include
disaster recovery and independent assisted living
applications. Nodal location awareness is clearly necessary
because a sensed event should be reported along with where
it happened.
Since for the scenario of many WSNs application GPS is not a
viable choice, localization protocols have been proposed
that provide nodal positions. In particular, it has been
observed that the concurrent deployment of inter-nodal range
and the angle of arrival (AoA) of the radio signal can be
effectively used for localizing nodes with remarkable
accuracy. While techniques for measuring the inter-nodal
range have been proposed that achieve accurate estimation,
determining the AoA is still widely uncharted territory in
WSNs. This research will investigate the design of a
physically and electrically small, low-power multiple
element antenna array that can provide unambiguous AoA for
RF signals.