Current Research Projects
The Tether-less Patient
The Tether-less Patient aims to endow patients with their mobility yet
provide healthcare professionals with similar or even improved monitoring
capabilities they would have with the patient confined to a hospital bed.
By monitoring patients as they go about their lives, the healthcare community
gains insight into the patients actual lifestyle, frees hospital resources
for other patients, and ultimately achieves lower costs. The patient returns
to work sooner, may avoid the nursing home or assisted living center longer,
and gains feedback to make changes to his or her lifestyle.
I'm particularly focusing on the communication infrastructure of these
systems. Where wireless carriers will tell you that their networks will
always be available, there cannot be any 100% guarantee, and that
margin of uncertainty requires our attention.
Simulating the Tether-less Patient Environment
At the moment, my work utilizes the java-based
Opportunistic
Network Environment Simulator. I've been responsible for some
contributions to the simulator, which I've made available
here.
A tether-less patient system or, more specifically, a delay tolerant
network is dependent upon the behavior, movement, and mixing of the nodes
participating in the network. The Mobile Usage Project is an attempt to
understand those properties given contemporary usage habits of smart phones,
which are in low-power, sleeping states most of the time. Most datasets
collect traces from devices that are always on and always looking for other
devices. How good is our network when the nodes are mostly asleep, though?
Virtual Repositories in Delay Tolerant Mobile Ad Hoc Networks
It is often difficult to distribute routing or position
information in sparsely connected or often partitioned ad hoc
networks due to infrequent communication contacts. It then becomes
much more difficult for a particular node to determine the location
of a desired destination in order to properly route a message to it.
We seek to add a distributed message store to the DTN layer
responsible for permanently storing messages whose destination
nodes’ locations are not currently known to some set of nodes in the
network. A globally known hash function maps each node to a fixed
position in the geographic area of the network. When a node has a
message for another node whose location is unknown, it hashes the
node’s ID to obtain some location in the geographic area, what we
refer to as the Virtual Repository for that node. The message is
then routed to this location with the hope that there exists some
set of nodes currently positioned near this location, on which we
impose the responsibility of collaboratively storing the message. It
then becomes the responsibility of the destination node to pull its
messages from its Virtual Repository whenever it becomes convenient.
Distributed Data Backup
Let’s say you’ve got some part of your hard drive that you’d like
to back up somewhere else in case your hard drive fails. If you’re
paranoid, you don’t want the other copy of your data to be
physically stored anywhere near your PC in case you spill orange
juice all over everything (or your PC is the target of a nuclear
strike)
