CEE 200: Structural & Geotechnical Engineering Seminar
UCLA Civil & Environmental Engineering Department
The Role of A-B-C (Artificial Intelligence, Building Information Modeling, Computer Vision)
in Proactive Project Controls on Construction Sites
Mani Golparvar-Fard, Associate Professor
University of Illinois at Urbana-Champaign
Achieving smooth flow of production in construction requires team-based planning and systematic avoidance
of waste through production control mechanisms. Over the past decade, production control theories such
as the Last Planner System have emerged that stabilize workflows by shielding the direct work from upstream
variation and uncertainty. While the benefits of these theories are well documented, yet their potential across the
life of a construction project is not fully achieved and the root-causes for this are not entirely understood. A
large body of empirical observations suggest that successful implementation of control mechanisms requires
dedicated facilitators and engages practitioners in a relatively deep learning process. Sustaining this level of
commitment for the duration of a project can be difficult and, in its absence, project teams may revert back to
traditional project control practices. While some attribute these barriers to the people and organizational processes
involved in implementing lean principles, yet there is a growing recognition that the functional aspects of
production control techniques need close re-examination to better understand, predict and analyze reliability in
performance, and preserve effective and timely flow of information both to and from the workface. To address
these knowledge gaps, this talk presents a new visual production management method that captures, communicates
and analyzes actual and potential construction performance problems. To ensure its implementation
does not take away from actual productivity, the method extends the value of 4D Building Information Models
(BIM) – used for constructability review- as a benchmark for performance. Likewise, it takes advantage from images
and videos frequently collected by project personnel via smartphone and drone cameras to visually document
actual performance. These images and videos are used together with new computer vision and machine
learning algorithms to continuously map the current state of production over project timeline. These 4D reality
models are integrated and compared automatically with the 4D BIM to expose actual waste and highlight potential
issues by forecasting reliability in project look-ahead schedules. The underlying machine learning algorithms
and methods used for various aspects of this method are presented in detail. Experimental results from deploying
this method will be shown to demonstrate that these models together with actionable data analytics on construction
performance can proactively support collaborative decision making that eliminates root causes of
waste. These models also provide visual interfaces between people and information that enable effective pull
flow, decentralize work tracking and facilitate in-process quality control and hand-overs among contractors.
Where: Mong Auditorium, Engineering VI
When: 12- 1 PM, Wednesday, April 24
Speaker: Dr. Golparvar is Associate Professor of Civil Engineering, Computer Science, and Technology
Entrepreneurship at University of Illinois at Urbana-Champaign. His work in the area of
computer vision and BIM for construction and infrastructure monitoring has been the recipient
of many awards including 2018 Walter Huber Research Prize from ASCE, 2017
National Top 20 Under 40 from ENR; 2016 ASCE Dan Halpin Award for Scholarship in
Construction and 2013 ASCE James Croes Medal for innovation in Civil Engineering; 2013
and 2010 CETI awards from FIATECH; 2017 Hojjat Adeli Best Journal paper Award for Innovation in Computing
from Wiley-Blackwell , 2012 best journal paper award from ASCE Journal of Construction Engineering
and Management and numerous best conference paper awards. Dr. Mani Golparvar is also co-founder of
Reconstruct, a Software-as-a-Service company that visually track progress, analyze productivity, and proactively
detect potential delays using predictive analytics; in turn empowering construction companies to keep
their projects on schedule and on budget. Reconstruct has been recognized by several industrial awards
such as 2016 World Economic Forum most innovative startup company and 2016 Innovation Award from
Turner Construction.