Summaries of each project are below.
Media
distribution on the net is maturing to become competitive with
mainstream
radio
and television, radically changing the industry. However, there are
critical technologies needed to make real time distribution scale
without
incurring astronomical bandwidth and server costs. At present
the only feasible
approach is to utilize a peer-to-peer architecture, where
every receiver acts
as a
relay contributing bandwidth back into the system. This comes with
its own
set
of hard problems, stemming from network asymmetry and the inherent
unreliability of end-user machines.
This project is developing an open platform for globally
scalable real-time
media distribution. The core technology is VidTorrent, a
distributed protocol
akin to BitTorrent but designed to work with real-time
streams, where one
cannot readily delay delivery. We are working towards making
the technology
ready for end-user deployment, while seeking avenues for
comercialization into
a viable global business.
Contact: Dava Newman
Forget today's 300-pound bubble space suit. Think about a
sleek biosuit that
will protect space travelers, enable them to exercise during
the six-month
journey to Mars, then power walk across the planet's surface.
That's the
30-year goal offered by Aeronautics and Astronautics
Professor Dava Newman The
core technologies, which include a compression skin suit and
electromechanical
devices, can be realized in the next few years in active
orthopedic devices
that enable stroke victims to regain mobility. Newman's group
has designed an
active orthotic device that propels a person walking. The
i-Team will focus on
shorter term earth applications for this technology.
Contact: Steve Massaquoi
Currently available robots require a structured operating environment for highly structured tasks. However, there is an increasing need for semi-autonomous and autonomous robots that are capable of safely working with humans in open environments with complex time-varying characteristics. The recent availability of inexpensive motors with high power to weight ratios combined with the dynamic control imposed by neuromimetic computer programs enables such robots to be managed simply, efficiently, and flexibly.
This project proposes to leverage the current advances in robotic navigation and intelligent manipulation to develop an inexpensive, high-performance, reliably stable robot with bipedal locomotion. The i-Team will explore a robot that is currently being constructed to demonstrate the feasibility of control via simulated spinal cord, brainstem, cerebellar and cerebral function. Once fully developed, these robots could be valuable for tasks currently performed by humans including: surveillance, search and guidance, the automation of manual labor tasks in open environments and work in mass production.
Contact: Christopher Thompson
Traditional flexible endoscopy has been limited to the
confines of the
gastrointestinal lumen, however, recent developments
involving transluminal
access to intra-abdominal structures hold the potential to
revolutionize
flexible endoscopy. Over the last few years various studies
have detailed
techniques involving intentional luminal breach with
endoscopic access to the
peritoneal cavity for diagnostic and therapeutic procedures.
Unlike standard
surgical approaches which require incisions in the abdominal
wall the Natural
Orifice Transluminal Endoscopic Surgery (NOTES) approach
avoids abdominal
incisions and may offer specific advantages in select patient
populations.
Brigham
and Women's Hospital and Harvard have patented a novel trochar for
transluminal access that will provide a sterile environment
enabling these
procedures to be safely translated to human use. This is in
the process of
being prototyped. Depending on the evolution of this new
field, this device may
be critical to procedural success and could be highly
utilized.
Contacts: Moungi Bawendi, Jonathan Halpert
Nearly every biotech and pharmaceutical company relies on
dyes or fluorescence
markers for standard lab test methods. Typically, the
byproducts of these
reactions are unstable and hard to detect. Colloidal
semiconductor nanocrystals
offer
unique emission and absorption properties that make them superior to
conventional dyes in terms of product stability and emission
spectrum. This
technology is also being explored for developing noveloptical
devices and
materials.
Our
project aims to evaluate the market potential of the various
applications of
nanocrystal technology, and identify the most compelling of
these applications.
Lastly, we aim take the actions required to access this
market.
Contact: Michael Stonebraker
A
fundamental problem for large enterprises revolves around converting
data from various representations inside the logic of business units
to one common representation in the enterprises central data
warehouse. Business analysts subsequently query this warehouse to
improve business processes.
While tools do exist to extract data from various locations and
transform them into a common format, such exercises are usually
complex and expensive. Moreover, todays modern business environment
necessitates frequent changes in the data structures of business
units. This makes such transformation exercises an ongoing issue.
The goal of this project is to create a next-generation data
transformation tool to facilitate these conversions.
The core technology is called Morpheus (the Greek god of
dreams, who could change his form and appearance) that consists of a
transformation tool and a repository that holds a large number of
transformations. Morpheus would be used as a browser tool, finding a
repository transformation that is "close" to the required
one. Then, the transformation tool is used to manipulate the
identified starting point in a drag-and-drop fashion quickly and
easily into the desired result.
Wireless networks are increasingly key to modern
communications, from cellular
phones to distributed sensors and emailing from the local
coffee shop. Yet, as
the number of wireless devices increase, and devices use
network connectivity
more often, the network is more heavily taxed and overall
throughput begins to
fall drastically in dense settings. This simple fact of
current network design
poses a serious threat to hopes of urban Wi-Fi networks and
similar emerging
applications.
This logjam of data packets can be broken by
"opportunistic listening" and
"opportunistic coding". Nodes in the wireless network snoop
on all
communications they hear over the wireless medium, and store
the heard packets
for a limited interval. The nodes also annotate the packets
they send to tell
their neighbors which packets they have heard. When a routing
node forwards a
packet, it uses its knowledge of what its neighbors have
heard to perform
opportunistic coding; the router node can intelligently
select and mix multiple
packets and send them in a single transmission. This more
efficient delivery
allows higher throughput.
Inserting needles for catheter placement into the human
body is one of the most common medical procedures performed on
millions of people annually in the United States alone—yet accidental
misdirection of the needle still occurs in hundreds of thousands of
patients, often causing severe complications. The central problem is
an inability to distinguish the type of tissue the needle encounters
as it is inserted into the body. Potentially devastating
complications include venous catheters entering an artery, or an
epidural needle penetrating completely through the intended epidural
space and puncturing sensitive dural tissue.
A
device that mechanically senses when a needle passes from a
high-resistance tissue region to a low-resistance region may address
these issues. By creating a probe equipped with an appropriate
mechanism, the catheter can react to the tissue it encounters and
facilitate precise understanding of its progress. This project aims
to develop and perform proof-of-concept experiments using prototype
devices for sensing tissue and tissue compartments.
Nobody questions the need for smaller, more efficient and
longer-lasting
portable power sources. In the United States alone, the
market for batteries
grows at six percent annually and is projected to hit $15
billion by 2009. But
today's battery technology largely depends on chemical
reactions — leading to
devices that degrade over time or possibly even malfunction
in dangerous “side
reactions,” as evidenced by the recent recall of millions of
lithium ion
batteries that could overheat in laptop computers.
One potential solution to this bottleneck is storing
energy in a dense network
of ultra-thin, microscopic filaments known as carbon
nanotubes. This project
focuses on the design and computer modeling of a single
nanospring-driven
generator and storage device. Detailed computer modeling
estimates will
determine the amount of energy that can be stored in a
nanospring device, and a
design and model for a generator and storage device with
optimal performance
will be created. Ultimately, prototypes may be built.
|