|
|
 |
|
|
|
|
MIRES
System Architecture
We proposed a mobile information resource
exchange system (MIRES) to implement resource exchange among
mobile users. In MIRES, mobiles are no longer just passive terminals;
by creating their Mobile Mirror Images (MMIs) and uploading
the resources from mobiles or importing others¨ existing data,
coordinated and controlled sharing of the resources can be realized
and achieved effectively through the MIRES model.
Optimizing Web2Mobile
System
Our model optmizes existing Web2Mobile
systems. This model focuses on how to make Web resources adaptive
to mobile devices, and how to reduce the network traffic through
optimizing the transmission. By reprocessing the text stream
and images which are the most frequently used ones at present,
the network traffic can reduce a lot on the application level.
Any multimedia resources containing text and image information
can adopt this kind of reprocessing. This opens up a door to
developing new and interesting applications which, on the other
hand, also require new service models to be in place.
Cooperative Image
Editing
Although traditional CSCW methods provide
many practical ways to realize image cooperatively editing,
their bandwidth requirement is relatively large, compared with
the narrow bandwidth of wireless network. Our proposed cooperative
image editing tool is more suitable for mobile phones. In our
tool, users need not update his/her local copy frequently. Instead,
he/she can finish a stage of editing work and then commit the
modification. Writing conflict can be resolved and merged automatically.
The latest one who commits the modification will be responsible
for judging the correctness of the merging.
Metadata-based
Mobile Resource Personalization
Under the framework of MIRES, Metadata
is used in the Bulletin Board to describe the content of shared
resources. A multi-level index is devised to cluster all metadata
in different levels. To realize personalization in MIRES, Bayes
estimation is used to serves as a filter, which lets end-users
receive only the results that they are actually interested in,
thereby saves the user from incurring unjustified cost on unwanted
resource items.
|
|
Mobile
Data Caching
We are studying a statistical caching
mechanism for mobile data access in cellular systems. This caching
mechanism makes use of prior knowledge (statistical data) to
predict the user mobility pattern. Based on the statistical
heuristic search, a cache maintenance algorithm is developed
with a computational complexity of O(Dln2D), which is acceptable
for realistic applications. Experimental studies show that our
caching mechanism can obtain good performance on caching mobile
data for cellular systems.
|
|
Data
Cache In Selfish MANET
We are studying a caching scheme and a
set of protocols to encourage the powerful mobile devices (strong
nodes) to provide data cache service to other weak nodes (e.g.,
mobile phones) in the MANET environment. We suppose that each
node in the MANET is greedy and self-interested, always trying
to maximize their own income when the market finally reaches
an equilibrium (which is called Nash equilibrium in microeconomics).
However, it is known that the maximization of the income of
individual nodes may not lead to the Pareto Efficiency or social
optimum (i.e., the maximization of the outcome of the game).
Through game theory, we have proved that our protocol design
can actually achieve social optimum. |
|
|
|
|
| |
![]() |
