CS Cache Interviews New Faculty Member
Nikolaus Correll
By Michael Main
Nikolaus and Olga Correll’s two children, Arthur and Tatiana, are growing up
multilingual in Boulder. There’s German and English for starters, but around the Correll
house, they may very well find themselves conversant with robots as well. That’s because
their father’s work in multi-robot systems has the potential to follow him home.
Nikolaus Correll joined the computer science faculty at CU Boulder two years after
completing his 2007 Ph.D. dissertation on swarms of miniature robots at the École
Polytechnique Fédérale Lausanne, Switzerland. Prior to joining us, he was a
During his first semester in Boulder, Nikolaus drew graduates and undergraduates into his
research projects, and is also collaborating with engineering faculty members Christopher
Bowman (Chemical and Biological Engineering), Robert McLeod (Electrical Engineering), and
In their recent research proposal, Nikolaus and his colleagues envision using ACM to
build artificial arms and hands that can sense and grasp much as a human does, both for
When your parents ask, what do you tell them about your work in multi-robot
systems?
As we use more and more robotic systems in our daily life, such as vacuum cleaners and
toys, and “Butler” robots emerging, I believe these systems’ “intelligence” can be
drastically improved by having robots working together and sharing information with
other “intelligent” objects in the environment. More recently, I am intrigued by taking
It sounds very futuristic, like something I would read in Isaac Asimov. Has the
future arrived? Is there any particular fiction that has influenced you?
I can’t deny that I was thrilled by the robotic fantasies of the 80s and 90s—my youth—
including Robocop, Terminator, Minority Report and more recently The Matrix, Iron
Tell us about some specific multi-robot applications that you and your students are
working on now: things that we might be able to come see in the next year or so.
Robots that guide customers in malls, perform delivery tasks in factories, or provide
services in restaurants have been demonstrated; I think we will see more of this in the
near future. We’re currently working on algorithms for robots that meet in tight
environments such as a factory floor or a crowded mall. They need to plan optimal,
What problems of scale do you anticipate when we go from a handful of robots to
thousands, or from thousands to billions of tiny robots?
Scaling problems can be mitigated by using decentralized algorithms that rely solely on
local communication and local information. While these algorithms are usually less
What’s the smallest hardware platform that you’re working with, and what
languages do you use to program it?
For our smart materials we are using Atmega 88s that have a footprint of 4×4 mm2. They
are programmed in C (in fact, the same gcc compiler that first-year students use
Does the notion of emergent properties arise in your work? Or perhaps self-
replication?
We see a lot of emergence, although more often due to bugs than by design. The problem
with the term “emergent” is that it is much less exciting as soon as you know the
What other challenges do you find most interesting from a standpoint of enjoying
your work?
I’m very excited about bringing people from other domains together to create
What were the key influences in you and your family deciding to come to Boulder?
Boulder is full of opportunities for robotics and interdisciplinary research, and I had a lot
of stimulating discussions during my, in visit. Also Massachusetts, I really missed the
As of yet, no swarms of robots have been spotted following Nikolaus back and forth
between the foothills near 5th Street and his office on the 7th floor of the Engineering