But this is not simply an academic exercise. Looking over the wireless horizon, Estrin sees wireless technology applications that sound as if they were lifted from the pages of a pulp sci-fi novel. In this emerging future, she says nanometer-sized sensors will track the path that pollutants take as they flow through soil, and "smart buildings" will sense and then adjust their bearings to survive earthquakes.
What do you think is going to be different about the wireless field
compared with today? Is it just a matter of smaller, faster, cheaper?
It's also a matter of smarter. It's the combination of sensing,
computation and wireless communication that seems to offer a more
qualitative change to the way we do things, the way we manage everything from agriculture, food transport and production, transportation, medical care, public health exposure studies, and contaminant transport monitoring.
What kinds of devices are you trying to build at the center?
We're focused on developing devices that can organize themselves into a network, repair themselves and manage their own power consumption. Our systems use actuators, which can do more than just look at something. They can take a sample, for instance. There are also sensors, which are little
devices capable of picking up visual and auditory information. These
things are sized on the scale of circuits on chips, so they can be
weaved into fabric or poured into concrete.
How far along are you with smart materials and smart structures?
At a device level, we are way ahead. There are sensors in cars, roads
and traffic lights. But what you don't have are the applications
distributed over a large region, like the ability to measure if
there's pooling of contaminants for instance. The ability to
exploit them in large-scale applications is missing.
Any answers yet on the scaling problem?
We're trying something similar to the biological systems. Every
stimulus that your nerve endings feel doesn't get sent back to the brain. We're trying to do the same thing. The scaling of the system means you want to build a distributed system where you locally process the signal, then collaboratively process the signal (using many computers placed in one location).
Where do you expect all this will be used?
Take the same idea to a rain forest to study microclimates, or a local
conservancy area to try and understand how it's being affected by urban development.
The Center is also exploring using wireless technology and seismic
sensors, creating a "smart building"?
In the long term, we are figuring out how to make this technology that, when poured into the concrete of a building, would enable (the building) to monitor its own health. For seismic applications, the
sensors on the building beams might be large.
You also have been working on what sounds like wireless trees. Can
you explain?
Think of a weather station shrunk down to a square inch with wireless
transmitters so that now you can measure what happens every cubic meter
on a forest floor or canopy. Scientists are looking for details about
microclimates. One question they have is why smaller climates don't see
the extreme changes going on around them, like intense rainstorms.
When it comes to predicting evolution and revolution in wireless
technology, how do you see the future? Will it progress in increments or do you expect big breakthroughs?
I am a researcher, not a futurist. I respect the latter but don't
consider myself particularly insightful in that realm. So my wimpy
answer is that it will be a combination. Some of the important
breakthroughs and the ones I am trying to enable will be
Internet/Web-like. They come about through important and communally
developed and used infrastructure. TCP/IP (protocols that allow
computers to talk to each other over long-distance networks) and the
Mosaic Web browser (which gave rise to Netscape) were breakthroughs in
that sense, but they couldn't be breakthroughs based on technical merits
alone. They were breakthroughs because they were picked up by a
community and built upon, and this amazing thing--that is the Internet--happened.
What are the biggest technology and regulation obstacles
to a wider wireless adoption?
Technology-wise, I think it will be making the systems
self-healing...How will we know what is working and what is not? We
have a hard time knowing what is going on in the Internet where
communicating is relatively cheap. How do we get enough information out of these systems to have confidence in them when the communication paths
are inherently narrow?
And regulation?
Regulation is not as much on my radar at this point. However, clearly,
when it comes to medical and public health and safety-critical
applications, there is the need for verification, validation,
robustness, and issues of liability that will definitely come into play.
Where is the most interesting research into wireless going on in
academia these days?
There is a lot going on so this list is incomplete. Just to name a
few: At the radio frequency and component level there is good work
going on to develop very small low-power radios at University of
Michigan's Wireless Integrated Microsystems Engineering Research Center
and the Berkeley Wireless Research Center. When it comes to wireless
systems, many campuses across the country have interesting and
innovative projects under way, including (work at) the University of
California, San Diego.
Any corporations?
Intel clearly gets it. They have three labs active in different areas
(of wireless): at the University of Washington, at the University of California, Berkeley and at
Carnegie Mellon University. And there is a tremendous amount of activity within Intel proper, including wearable computing sensor networks.
How long do you think it will take for some of that to make its way
out of the lab and into commercial products?
It's hard to say in this economy. But I am confident that we will
see some commercialization over the next few years.
