News

IACAT supports six collaborative projects

Published Date: Aug 13, 2013

Six collaborative projects will bring together the expertise of University of Illinois faculty and the skills of staff at the National Center for Supercomputing Applications thanks to support from the Institute for Advanced Computing Applications and Technologies.

The projects selected for individual and collaborative fellowships for 2013-2014 are:

  • Energy Performance Augmented Reality Models for Building Diagnostics Using a Hybrid of RGBD/Thermal Cameras and CFD Models. Mani Golparvar-Fard (Civil and Environmental Engineering) aims to automate the process of generating accurate and complete Energy Performance Augmented Reality (EPAR) models for easy and quick building diagnostics through advances in computer vision and construction management. The purpose is to provide energy auditors, facility managers, and homeowners with information they need to easily make retrofit decisions that directly impact the energy efficiency of existing buildings.
  • Visualizing Topic Models about African American Women’s Everyday Experiences and Standpoints. Ruby Mendenhall (Sociology, African American Studies, and Urban and Regional Planning) will work with NCSA visualization specialist Mark Van Moer and Michael Simeone from the Illinois Informatics Institute to apply text mining techniques, topic modeling, and data visualization to determine key concepts and relationships in more than 8 million periodicals in JSTOR from as far back as 1923 and over 5 million books and newspapers in the HATHI Trust Digital Library. The goal is to capture the nuances of African American women’s experiences and their efforts to negotiate and maximize resources in their everyday lives.
  • Subsampling Methods in Modeling of Big Data in Geosciences. Ping Ma (Statistics), Jong Lee (NCSA) and Shaowen Wang (Geography and Geographic Information Systems) will tackle big data challenges by applying their “smart algorithms” to small, representative subsets of data, extracting the relevant information without the need for massive number-crunching power.
  • Autonomous Vision-based Progress Monitoring of Building and Infrastructure Construction Projects. Mani Golparvar-Fard (Civil and Environmental Engineering), Timothy Bretl (Aerospace Engineering) and Derek Hoiem (Computer Science) plan to use aerial robots—quadcopters—with onboard cameras to gather image and video data of construction progress. Their ultimate goal is to automate the process of large-scale building and infrastructure construction monitoring through computer vision-based 3D reconstruction methods and Building Information Modeling, and provide owners, contractors, subcontractors, and tradesmen with the information they need to easily and quickly make project control decisions. The collaborators hypothesize that both construction cost and delivery time can be significantly reduced with tools that better characterize the extent to which construction plans are being followed.
  • Fractal Patterns in Fracture and Damage Phenomena. Preliminary research by Martin Ostoja-Starzewski (Mechanical Science and Engineering) and Seid Koric (NCSA) indicates that the fractal dimension can be used to estimate the level of fracture and distributed cracking for damage assessment of materials. This offers a rapid and inexpensive alternative to cumbersome strain gages and unwieldy stress measurements in a vast range of applications: safety and reliability of bridges, buildings, machine parts, aircraft, etc. Using NCSA’s computational power, the collaborators plan to introduce spatial material disorder/randomness into models of elastic-plastic-brittle materials, and to study the growth of multiscale systems of interacting cracks.
  • Cyber Tumors for Predicting Cancer Behavior in Human Patients. Rohit Bhargava (Bioengineering), Narayan Aluru (Mechanical Science and Engineering), Andre Kajdacsy-Balla (Pathology, UIC), Partha Ray (Carle Foundation Hospital) and Kenton McHenry (NCSA). The goal of this project is to develop a virtual model of the human breast. Properties of the developed model can be tailored to a specific patient; a virtual tumor can be grown computationally and the resulting optical and chemical imaging data will be predicted.

For more information about the IACAT Fellows program, see http://iacat.illinois.edu/faculty_fellows.