IEEE Spectrum November, 2017 - 6

SPECTRAL LINES_

11.17

Need to Build a Better Hospital?
Use Chip Design
The construction industry has gone undisrupted
for nearly a century. It's time

eepak Aatresh, an electrical and computer engineer from India,
joined Intel in 1989 as a chip designer-after reading an article
I wrote for IEEE Spectrum on Intel's project to build the first
million-transistor chip! After seven years at Intel, he moved on to
communications hardware companies. ¶ Then, in 2008, bitten
by the startup bug but without a particular startup idea in mind,
he joined Artiman Ventures as an entrepreneur-in-residence. He
thought he'd try to find ways to reduce energy waste and improve
energy efficiency. ¶ Although he didn't know much about it at the start,
he zeroed in on building construction because buildings consume huge
amounts of energy. He taught himself how construction works, from site
clearing to excavation to putting up a building. As part of his research,
he watched a time-lapse video created from images taken by a camera
mounted at a construction site, in which more than a year of work flicked
by in minutes. ¶ "That's when it hit me," he says. "It's just like we make
silicon chips-first we etch away, and then we deposit layers. We have the
same end process." ¶ He knew that the two industries had very different
design processes. But then he thought, why should they? Why shouldn't
buildings be designed the same way chips are designed, using automated
tools, circuit libraries, simulation, and formal design verification? ¶ With
Zigmund Rubel, he founded Aditazz to try to make a go of it. "My instinct was
to make it a software play," Aatresh says, meaning to build the design tools
and sell the software. His software works in essentially the same way that
chip design tools do. It has libraries of different space patterns (the design
elements) and other rules-for example, the spatial elements that make
up, say, an ultrasound room. "We can algorithmically adjust the design for

NOV 2017

NORTH AMERICAN

SPECTRUM.IEEE.ORG

the volume of patients or size of the footprint." Essentially, Aatresh says, "we create a functional description of a building,
and use a compiler to build the design."
Before going to market, Aditazz had
to prove that its algorithms worked. The
best way to do that was to start out as a
service provider, bidding on contracts
and designing great buildings using
the new tools. (This approach is being
embraced by other entrepreneurs; see
my September View From the Valley post
"Starting a Robotics Company? Sell a Service, Not a Robot.")
Aditazz decided to focus on the design
and construction of hospitals and other
health care facilities, reasoning that these
are particularly difficult design problems that could benefit significantly from
being approached with better tools.
So the company entered Kaiser Permanente's "Small Hospital, Big Idea" competition. The challenge was to design a
100-bed hospital that could be duplicated
around the country. About 300 firms
applied. Aditazz was one of two winners.
Kaiser had initially planned to build several dozen
hospitals based on the winning design, but due to
uncertainty surrounding the U.S. Affordable Care
Act, the project has not gone forward. The win still
gave Aditazz credibility, however, and the firm is
now focusing on China, where it is using its technology to redesign a cancer hospital and research lab.
"Our approach," says Aatresh, "started [by looking
at] the incidents of cancer in their catchment area,
to determine how many patients might need to be
treated" at the Chinese site. As a result, Aditazz ended
up with fewer beds than in an earlier design and a better match between the patient population and the
diagnostic, treatment, and outpatient areas, shrinking the overall size of the proposed building by about
30 percent.
The chip designer in him can't help but be pleased
with these results. "From a semiconductor perspective, if you can shrink the die size by 30 percent, that's
huge," Aatresh says. "Power goes down, frequency
goes up, cost goes down. You improve quality and
reduce cost." Very nice indeed. -Tekla S. Perry
An extended version of this article appears in our View
From the Valley blog.
↗ POST YOUR COMMENTS at http://spectrum.ieee.org/spectrallines1117

PHOTO-ILLUSTRATION BY

Edmon De Haro

https://www.spectrum.ieee.org/view-from-the-valley/at-work/start-ups/starting-a-robotics-company-sell-a-service-not-a-robot https://www.spectrum.ieee.org/view-from-the-valley/at-work/start-ups/starting-a-robotics-company-sell-a-service-not-a-robot http://www.kaiserpermanentehistory.org/latest/what%E2%80%99s-the-big-idea-seeking-new-concepts-for-the-ideal-health-care-setting/ http://www.artiman.com/ https://www.aditazz.com/ https://spectrum.ieee.org/view-from-the-valley/at-work/start-ups/how-chip-design-can-teach-us-to-build-better-hospitals http://spectrum.ieee.org/spectrallines1117 http://SPECTRUM.IEEE.ORG

Table of Contents for the Digital Edition of IEEE Spectrum November, 2017