Dear Members and Friends of CITRIS,

Looking back on 2003, it has proven to be a year of tremendous progress for CITRIS. Our researchers and their projects have garnered national and international attention in both the media and scientific communities, the scope and interdisciplinary nature of CITRIS has expanded to include eight unique application areas covering dozens of disciplines, and our support from industry and government has grown substantially. As we approach the mid-point of our third year, we continue in our pursuit to become an even more collaborative and innovative organization, with the kind of impact on society embraced by the original vision for CITRIS.

OCTOBER/NOVEMBER HIGHLIGHTS:

• In October CITRIS held its semi-annual Corporate Sponsor Day, this time at UC Santa Cruz. The two-day event was a resounding success featuring over 15 speakers, 40 student posters and 125 attendees representing industry, government, and academia. For full event highlights and to download event presentations and posters, please visit citris.citris-uc.org/events/spotlight/fcmday_1003.shtml.
• Also in October CITRIS hired Claudia Bruno as a Grants Administrator. Claudia is responsible for managing awards granted to the CITRIS Director and Principal Investigators. Claudia can be reached via email at cbruno@eecs.berkeley.edu.
• CITRIS continues to receive generous support from our Founding Corporate Members and Associate Corporate Members. Most recently we received contributions from Infineon Technologies and Microsoft Corporation totaling over $730K. The funds received from Microsoft Corporation will be used specifically to support the ICT4B project that is dedicated to developing a scalable enabling IT infrastructure for developing regions.
• Also noteworthy is the CITRIS co-sponsorship of the NSF Lake Tahoe Workshop on Collaborative Virtual Reality and Visualization. The three-day event focused on collaborative virtual reality and its application to large-scale data set and sensor network data visualization.

On behalf of CITRIS I want to wish you all the best during this holiday season!

Professor Ruzena Bajcsy
Director
Center for Information Technology in the Interest of Society

CITRIS Awards, Honors, & News

December 2003

• 12.03.03
  CITRIS researcher David Culler named among Scientific American top 50 innovators of the year.
  Related Article:
  Mercury News

November 2003

• 11.25.03
  CITRIS co-sponsors tradeshow that showcases life-simplifying inventions by UC Berkeley engineering students.
• 11.18.03
  UC Davis CITRIS researchers awarded $5 million for cybersecurity study.
• 11.17.03
  CITRIS researchers probe data avalanche.
  Related Article:
  Where is your data headed?
• 11.12.03
  CITRIS researchers try to measure the amount of information that humans create.
  Related Article:
  USATODAY.com
• 11.03.03
  On Demand Manufacturing used by CITRIS researchers in developing robotics and electronics.
• 11.03.03
  CITRIS researchers build a crash-test Internet that will assess vulnerability.

<!-- InstanceBeginEditable name="Feature1Title" -->CITRIS Q&A <!-- InstanceEndEditable -->

<!-- InstanceBeginEditable name="Feature1Summary" -->Interview with Bernd Hamann, co-director of UC Davis's Center for Image Processing and Integrated Computing<!-- InstanceEndEditable -->

by <!-- InstanceBeginEditable name="Feature1Author" -->David Pescovitz<!-- InstanceEndEditable -->

UC Davis's Associate Vice Chancellor for Research, Bernd Hamann, is knee-deep in data. As a CITRIS researcher and co-director of UC Davis's Center for Image Processing and Integrated Computing (CIPIC), Hamann develops novel methods to transform massive amounts of numerical data generated by computers, imaging technology, and sensor networks into meaningful, and often beautiful, computer graphics. Through Hamann's novel data exploration methods, scientists from myriad disciplines gain a clearer picture of what's happening outside in the real world, and inside our own bodies.

Why is the development of new data visualization methods important to CITRIS?
Hamann: CITRIS is a driver for the development and deployment of large sensor networks that collect vast amounts of data about everything, from the structural integrity of bridges during an earthquake to a building environment after the outbreak of a fire. People will instantaneously want to see pictures representing the recorded data so they can analyze it easily. Our role is to develop innovative visualization technology to help comprehend massive data floods rapidly.

Besides information from sensor networks, what other kinds of data do you visualize?
The "Informatics of Human and Monkey Brain Atlases" is a collaboration between the UC Davis Center for Neuroscience and CIPIC. One of the components is research directed at the effective visualization of human and monkey brain data obtained from, for example, magnetic resonance imaging. A major question is how to define, then characterize mathematically, and ultimately visualize abnormalities in diseased, abnormal brains. A "brain atlas" can be thought of as a reference model, representing a prototypical brain structure. Using our interactive data exploration tools, a neuroscientist can, for example, re-slice a particular individual's brain and eventually compare it in qualitative and quantitative ways with a reference brain atlas to discover areas that may diseased. Schizophrenia is one of the diseases that Edward Jones, the principal investigator of this project, is studying.

What is the biggest technical hurdle you face developing new visualization technology?
One major challenge is how to allow scientists to access and browse vast amounts of data remotely over high-speed networks. We would like scientists who are remote from the data to be able to visualize and interact with that data with minimal delay so that they can explore it in real-time.

For more information:

Bernd Hamann's home page
http://graphics.cipic.ucdavis.edu/~hamann/index.shtml

The Brain Atlas Project
http://nir.cs.ucdavis.edu/index.jsp

Center for Image Processing and Integrated Computing (CIPIC)
Visualization and Graphics Research Group
http://graphics.cs.ucdavis.edu/

<!-- InstanceBeginEditable name="Feature2Title" -->CITRIS Feature Article <!-- InstanceEndEditable -->

<!-- InstanceBeginEditable name="Feature2Summary" -->The Real World as One Giant Database<!-- InstanceEndEditable -->

by <!-- InstanceBeginEditable name="Feature2Author" -->David Pescovitz<!-- InstanceEndEditable -->

The real world is becoming a database. Wireless sensor networks are being tested to monitor everything from a skyscraper's structural health to the sensitive ecosystems of endangered sea birds. The question though is how do you extract the valuable nuggets of information from a sensor network's firehose of raw data?

That's the idea behind TinyDB, a database system developed by CITRIS in collaboration with the Intel Research Laboratory Berkeley. TinyDB is part of an ongoing CITRIS effort to deploy wireless networks of "motes," tiny sensors only a few cubic centimeters in size that collect light, temperature, humidity, and other data about their physical environment. The data is then relayed from mote to neighboring mote until it reaches its desired destination for processing. For example, in one ongoing project, sensors in the burrows of endangered sea birds on Great Duck Island, near Acadia National Park, Maine, monitor the environmental factors affecting the creatures' comings and goings. In another effort, motes embedded behind a building's walls help diagnose seismic stability.

Essentially, TinyDB transforms these diverse kinds of sensor networks into user-friendly virtual databases that scientists can query in a familiar database language.

"Thinking of sensors as a collection of facts that you query rather than as a network of individual computers that you task frees you from the dizzying job of programming each one," says Berkeley computer science professor Joseph M. Hellerstein, also the current director of the Intel Research Laboratory Berkeley. "That way you can spend your time determining what you want to know from the network."

Hellerstein co-developed TinyDB with his former PhD student Samuel Madden, UC Berkeley computer science professor Michael J. Franklin, and Intel Research scientist Wei Hong. Through a graphical user interface, the software describes what sensor readings are available in a given network. Meanwhile, TinyDB's declarative query language enables the user to describe the desired data without having to tell the software how to acquire that data. For example, a scientist studying Great Duck Island might like to know the average number of birds in burrows during a certain time of day. The query is then sent to the TinyDB query processor pre-installed on each mote.

As it turns out, providing the motes with some data processing capabilities also saves power, a premium on the tiny devices. Transmitting data burns out the motes' batteries very quickly. The less data transmitted, the longer the motes live.

"You need to push the data reduction and summarization inside the network so that these things are communicating higher level concepts to each other rather than all of their readings," Hellerstein says.

TinyDB also acts as the network physician--tracking neighboring motes, maintaining how messages are routed across the network, and ensuring that every mote is reliably delivering its data.

"There's a rich tradition of database development at UC Berkeley," Hellerstein says. "And through CITRIS, we're now applying that computer science to the physical world."

For more information:

Intel Research Berkeley
http://intel-research.net/berkeley/

TinyDB Software Release
http://telegraph.cs.berkeley.edu/tinydb/

Joseph M. Hellerstein's home page
http://db.cs.berkeley.edu/~jmh/

Michael J. Franklin's home page
http://www.cs.berkeley.edu/~franklin/