CUDA Spotlight: Pete Willemsen
By Andrea Hutchison and Calisa Cole, posted August 13, 2012
GPUs for Green Urban Planning
This week's Spotlight is on Dr. Pete Willemsen, Associate Professor in the Department of Computer Science at the University of Minnesota Duluth.
Pete's research interests are centered on simulation and interaction within virtual environments (SIVE).
With several talented undergraduates and graduate students, he runs the SIVE Lab at the University of Minnesota Duluth.
This interview is part of the CUDA Spotlight Series.
NVIDIA: Pete, what is Green Urban Planning?
Pete: Green Urban Planning can take on variety of meanings. For instance, constructing buildings to minimize energy use, strategic placement of buildings within an urban space to reduce pollution exposure, or adding green infrastructure such as parks and green rooftops to urban landscapes to lessen heat buildup in an area.
However, the complex interactions between these greening ideas within the larger scale urban environment in which they have been applied are not well understood. Understanding the complex interactions between urban form, structure and environmental surroundings is the objective behind the GENuSiS project.
NVIDIA: Tell us about GENuSIS.
Pete: GENuSiS (Green Environmental Urban Simulations for Sustainability) is a highly collaborative and multi-disciplinary project with colleagues from the University of Utah's Mechanical Engineering, Atmospheric Sciences and School of Computing Departments and the University of Minnesota Duluth's Department of Computer Science.
Team members include Dr. Eric Pardyjak, Dr. David Johnson and Dr. Rob Stoll. The project is supported by the National Science Foundation.
Through the use of GPU-based simulation we hypothesize that there are urban structures and well-placed green infrastructure that can help to minimize energy use while also minimizing air pollution exposure. We hope our results can help guide future Green Urban Planning projects with knowledge that can maximize sustainability efforts.
Ultimately, our goal is to have our tools be the guidance for designing cities that are aesthetically pleasing yet present healthy situations within the urban spaces for peoples' active lifestyles.
NVIDIA: Why is there a need to focus on making cities more sustainable?
Pete: Much of the world's population currently resides in urban areas. In fact, based on information from the United Nations, just over half of the world's population lives in cities. That is expected to keep increasing. As such, urban areas account for a large percentage of overall energy use and waste production.
View of Tokyo, one of the world's most densely populated cities
Additionally, many cities are affected by air quality problems. There is a critical need to understand the complex interactions between urban form and the environment so that we can make well-informed decisions that positively impact these urban centers for the long term. The effects of this understanding can have positive ripples on sustainability and health.
"There is a critical need to understand the complex interactions
between urban form and the environment."
NVIDIA: What is the role of GPU computing in this project?
Pete: The use of GPU computing is affording us the parallelism necessary to run pollution dispersion and radiant heat energy balance simulations fast and on the design scales necessary for assisting with urban design decisions.
Our analyses require more than a single simulation for a specific urban space to understand the potential effects of the local environment. Rather, we are interested in iterating over fairly large decision and design spaces, running lots and lots of simulations to try to understand what urban configurations appear to mitigate pollution or lower energy use. As an example, consider attempting to minimize street level pollution, where human exposure is greatest.
Screen shot showing surface temperatures of Salt Lake City, created with GPU-based tools
We have used a variety of GPU-assisted coding approaches, including OpenGL GLSL, CUDA and even NVIDIA's OptiX ray-tracing API. Using CUDA and OptiX has allowed us to get a variety of students with different backgrounds working to parallelize our simulation routines and models. In particular, OptiX and CUDA have provided components necessary for a relatively straightforward mechanism for simulating radiant energy heat transfer in urban spaces rapidly and at a small scale.
We are also integrating the BOINC project into our optimization system to allow others to assist with solving these challenging problems on their own GPUs. We hope to have that available and online by the end of summer. Keep posted to our web pages for more details.
NVIDIA: How is your work impacting the way urban planning is done?
Pete: We really want urban planners and architects to ask questions like the following: "How do design choices – such as building placement, height, width, structure and/or orientation - impact how pollutants are dispersed, especially when the local urban context is taken into account?"
To answer these questions, we use various optimization schemes to work over quite large problem domains, running thousands to tens of thousands of simulations, attempting to locate and refine our ideas for which configurations help minimize pollution. The configurations can be provided to architects, urban planners and others to help inform them about structure and layout in the urban environment that can help to minimize pollution.
As another example and major focus of our current work, consider how green infrastructure can impact an urban environment. Green rooftops, parks and other greenscaping can help to change the urban fabric to be more conducive to reducing energy usage, water consumption and heat absorption as well. The interactions between urban structure and green urban scape are complex.
The “green roof” of Chicago City Hall. Photo DOE/NREL, by K. Scholz-Barth
Again, we would like planners to ask questions centered on helping them understand the impact of potential green structure on the urban space and how it affects the larger surrounding urban environment at larger scales. Our GPU-based simulations and tools are designed to help provide guidance in these situations.
NVIDIA: What's in store for future urban environments?
Pete: I think the future is good. We are at a point in time where much scientific knowledge is being generated about how urban environments work, how they should be created, managed and sustained.
It would be beneficial for urban planning to be informed by science, at least more than it might have been in the past. Moreover, producing tools that are amenable to different design strategies can help designers get easy access to science and engineering.
GPU hardware and software development is an important element in getting these tools to the planners, designers and decision makers. The outcome is that this new knowledge can help to positively affect our decisions regarding these spaces and the millions of people who live in and around them.
http://envsim.d.umn.edu/ (Results, images, video; list of project team members)
http://www.d.umn.edu/~willemsn/ (Pete Willemsen)
http://mech.utah.edu/~pardyjak/ (Eric Pardyjak)
http://www.eng.utah.edu/~rstoll/ (Rob Stoll)
http://www.cs.utah.edu/~dejohnso (David Johnson)
Bio for Pete Willemsen
Pete Willemsen is an Associate Professor in the Department of Computer Science at University of Minnesota Duluth. His research interests include perception and computer graphics, and simulation and environment representation for virtual environments. He has a Ph.D. from the University of Iowa.
willemsn (at) d.umn.edu
This material is based upon work supported by the National Science Foundation under Grant Nos. 1134580, 1133590 and 0828206. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.