Envision of an Integrated Information System for Projectdriven Production in Construction

Antunes, R. and Poshdar, M. (2018). “Envision of an Integrated Information System for Project-Driven
Production in Construction.” In: Proc. 26th
Annual Conference of the International. Group for Lean
Construction (IGLC), González, V.A. (ed.), Chennai, India, pp. 134–143. DOI: doi.org/10.24928/2018/0511.
Available at: www.iglc.net
ENVISION OF AN INTEGRATED
INFORMATION SYSTEM FOR PROJECT-
DRIVEN PRODUCTION IN CONSTRUCTION
Ricardo Antunes1 and Mani Poshdar2
ABSTRACT
Construction frequently appears at the bottom of productivity charts with decreasing
indexes of productivity over the years. Lack of innovation and delayed adoption, informal
processes or insufficient rigor and consistency in process execution, insufficient
knowledge transfer from project to project, weak project monitoring, little cross-
functional cooperation, little collaboration with suppliers, conservative company culture,
and a shortage of young talent and people development are usual issues. Whereas work
has been carried out on information technology and automation in construction their
application is isolated without an interconnected information flow. This paper suggests a
framework to address production issues on construction by implementing an integrated
automatic supervisory control and data acquisition for management and operations. The
system is divided into planning, monitoring, controlling, and executing groups clustering
technologies to track both the project product and production. This research stands on the
four pillars of manufacturing knowledge and lean production (production processes,
production management, equipment/tool design, and automated systems and control).
The framework offers benefits such as increased information flow, detection and
prevention of overburdening equipment or labor (Muri - 無理 ) and production
unevenness (Mura - 斑), reduction of waste (Muda - 無駄), evidential and continuous
process standardization and improvement, reuse and abstraction of project information
across endeavors.
KEYWORDS
Lean construction, SCADA, machine learning, LiDAR, BIM.
INTRODUCTION
In manufacturing, the operation is constantly monitored by the supervisory control and
data acquisition (SCADA) system. The system monitors, gathers, and processes real-time
1
The University of Auckland, Auckland, New Zealand, +64 20 40 12 4793, rsan640@aucklanduni.ac.nz
2
Lecturer, The Auckland University of Technology, Auckland, New Zealand, +64 921 9999 ext. 8956,
mani.poshdar@aut.ac.nz

Envision of an Integrated Information System for Project-Driven Production in Construction
Enabling Lean with IT 135
data from devices such as sensors and cameras, recording events into a log file and/or
displaying the operational information to local and/or remote locations through human-
machine interface (HMI) software. Because the information is available as soon as
possible corrective actions can be taken almost immediately. With the current
advancements in computing, intelligent models can also run in real time to detect future
issues supporting preventive actions. Despite SCADA systems and automation being
standard production tools in manufacturing their use in construction is minimal and
limited to isolated equipment.
In manufacturing, the production moves from machine to machine, worker to worker,
or a combination of both. The route of production is fixed (Antunes and Gonzalez 2015;
Hayes and Wheelwright 1979). Thus, the positions of sensors and actuators are fixed and
planned according to the production routes and its flow. Once set, the positions only need
to be modified if the production routes change. In construction production routes are
flexible. “Jobs arrive in different forms and require different tasks, and thus the
equipment tends to be relatively general purpose (Hayes and Wheelwright 1979).” Some
production routes will only exist long after the beginning of the project by the time that
others would be extinct. Construction must then rely on general purpose sensors that, as
the equipment, can be used in different applications through the project life-cycle, often,
requiring those also to be mobile. Hence, traditional instrumentation (and sensor
positioning) used in a manufacturing SCADA systems do not work in construction, as the
instrumentation must be mobile.
Building Information Modeling (BIM) can be considered as the closest system to a
SCADA applied in construction. BIM is the only system in construction that may contain
the production layout. However, BIM focuses mostly on production planning (Nederveen
and Tolman 1992; Rossini et al. 2017). The monitoring and control are still performed
manually regardless of the use of BIM. The production aspect of BIM, as well the general
industry, relies on primitive project management practices such as critical path and Gantt
charts [the latter neither being the first nor the most sophisticated production tracking
approach (Antunes 2017; Wesolowski 1978)]. These obsolete practices have been
abandoned in the industries with high productivity, such as information technology.
Construction occupies the bottom of productivity charts even showing negative indexes
of productivity over the years (National Society of Professional Engineers 2014). Some
common issues are lack of innovation and delayed adoption, informal processes or
insufficient rigor and consistency in process execution, insufficient knowledge transfer
from project to project, weak project monitoring, little cross-functional cooperation, little
collaboration with suppliers, conservative company culture, and a shortage of young
talent and people development (Almeida and Solas 2016).
Although much work has been done on implementing information technology and
automation in construction their application on an integrated flow of information is sparse.
This paper proposes a framework based on the current literature and technology to
implement automatic monitoring and control for construction management and operations
that could be useful to address the biggest issues of production in construction.
Conjointly, this research uses four pillars of manufacturing knowledge and Lean

Ricardo Antunes and Mani Poshdar
136 Proceedings IGLC-26, July 2018 | Chennai, India
production: production processes, production management, equipment/tool design, and
automated systems and control. The goal should be achieved by both top-down and
bottom-up approaches. The top-down approach will tackle the production system
collecting information about the construction environment and its changes. The bottom-
up approach will analyze the worker’s activity. By using smart-tools, embedded hardware,
Internet-of-things (IoT) and tracking the effort of labor can be measured and related to
project progress. The two approaches are stitched together by a machine learning engine
which makes sense of the data and the production theory comparing what has been done
with the plan provided in the BIM model.
TECHNOLOGY
BUILDING INFORMATION MODELING
BIM is a powerful, yet ‘promising’ tool for the design and construction industries.
‘Promising’ standing for both what it can do at the present and in the future. BIM is still
seen as a new technology in construction despite the increasing adoption and awareness
of BIM over the years (McGraw Hill Construction 2012; National BIM 2017).
The concept of BIM can be pinpointed back to the year of 1962 when Engelbart
presented a hypothetical description of computer-based augmentation system (Engelbart
1962). The application of computational solutions in construction was researched a bit
later
(Eastman 1969, 1973). The research focused on the automated space planning using
artificial intelligence in the bi-dimensional realm. The term ‘Building Information
Management’ appeared 30 years later (Nederveen and Tolman 1992) while the first
commercial implementation using this term is credited to ArchiCAD (successor of Radar
CH from 1984 for the Apple Lisa Operating System). Historically, it is important to note
that BIM did not derive from bi or tri-dimensional CAD. BIM (concept) is contemporary
of CAD development. Nevertheless, BIM as a tool built upon CAD three-dimensional
design tools for building modelers, which was a fully developed graphical tool for
building modeling available at the time.
The manufacturing industry explored further benefits of the tool besides graphical
modeling, in particular, parametric information technology tools (Autodesk 2002). Forms
in CAD drawings evolved to objects with the development of object-oriented
programming languages and their implementation to CAD systems in the early 1990’s.
Objects can bear graphical and non-graphical information bringing advances in both areas.
From a graphical perspective, instead of drawing elements, one by one the user could
design them separately and insert and reuse objects in the desired location. The group of
lines, forms, and surfaces is interpreted as a three-dimensional geometrical model of the
element it represents, for instance, a door or a window. The non-graphical perspective
gives meaning to that object. The object contains multiple graphical information, such as
the drawings of the door opened and closed. The object can have parts, and these parts
can be of different materials with different properties. The objects may also contain
production information attached, such as cost, labor, schedule, and effort what will give

BIM means to serve as a planning tool. Furthermore, changes required can be done in the
element and automatically replicated where it has been used rather than laborious one by
one changes. Overall, the reusable objects can bare more details and libraries of objects
could be developed and shared.
VIRTUAL REALITY AND AUGMENTED REALITY
Both virtual reality (VR) and augmented reality (AR) make use of 3D models to create a
scene in which the user can freely observe and/or interact with the models. What set these
technologies apart is how they use the background where the objects lie. Virtual reality
fully immerses the user providing a background to the environment. The user has the
perception of being physically present in a non-physical world. Conversely, augmented
reality utilizes the real environment as the background to project the models. The user is
partially immersed. Each one has different applications. Using 3D models, VR can
display a fictional scenario, for example, a functioning underground subway station even
before excavations begin. AR requires a background, thus, at least part of the station must
be in place. That is due to the fact that AR requires the recognition and tracking of
environment specific points for object placement. Both VR and AR are useful as HMI.
LIGHT DETECTION AND RANGING
Light Detection and Ranging (LiDAR) is a remote sensing method, which uses light
reflection to measure distances. The emitter shoots the light (ultraviolet, visible, or near
infrared) which is reflected and then captured by the receptor. As the speed of light, c, is
known, the time between emission and reception, t, is used to calculate the distance, d,
from the emitter to the reflector and back to the sensor, i.e., t=2d/c. An Global
Positioning System (GPS) receiver and an Inertial Measurement Unit (IMU) provide the
absolute position and orientation of the sensor. Thus, it is possible to calculate the
position coordinates of the reflective surface. One implementation of LiDAR consists of a
vertical array of emitters mounted on a rotational plate creating a linear scan that sweeps
the surroundings at each rotation. The result is a cloud of points which describes the
environment around the sensor. Despite the fact that the cloud of points provides accurate
measurement; the data does not identify objects. Basically, this cloud consists of x, y, and
z coordinates of each point. Making sense of what a group of points is often is a manual
task. Another limitation of LiDAR scans is the ‘shadowing.’ Because the technology
relies on reflection, it can make sense of lies behind a reflective surface or at the non-
reflective surface, such as water. The shadowing effect can be eliminated by scanning the
environment from different locations and thus overlapping cloud points [once the LiDAR
scans are almost ever combined with Global Positioning System (GPS) and inertial
measurement units].
LiDAR has been integrated to BIM aiming to identify defects (Wang et al. 2015).
That happens by comparing LiDAR measurements against BIM model specifications.
Deviations out of determined bounds are then identified as defects (Muda – Level II). A
quadcopter (any other carrier is possible, such as an aquatic or terrestrial drone or even a
backpack) inspects the site using LiDAR (inspection may also be considered Muda of
over-processing given the idea that the task should be done correctly instead of being

inspected for the approval). In this approach the defect flag rises without human
interaction, it however does not characterize a real-time system. The first reason is that
the defect will only be detected when (and if) the drone finds the issue, not at the time the
defect occurred. The second is that real-time systems require a timely response to the
event. A response out of the time-frame often results in catastrophic failure. The response
to the identified defect is not time-dependent. A real-time system depends on both the
logical result of the event and the physical instant features (Kopetz 2011). For instance,
the quadcopter drone moves forwards when detects on obstruction in its trajectory (event).
The trajectory correction (response) must happen in a timely manner otherwise the drone
will crash.
IMAGE
Image analysis can be an important tool in construction with several applications during
the project life-cycle. A simple application can the defect inspection, where the inspector
reports the non-conformities by taking pictures of the items out of specification and
which will feed a punch list to be addressed by the contractors (Muda of rework). The
pictures are used as evidence of the status of the non-conformities detected.
Additionally, because special cameras/lenses/sensors can capture infra-red and
ultraviolet, which are invisible to the human eye, the collected information can be used
for evaluating thermal and light insulation. Depending on what the cameras are mounted,
they can provide visual information from specific angles that are known to be dangerous
for human inspection (e.g., in confined spaces), or even impossible (e.g. for the pipelines).
The combination of multiple images provides even more information. Aerial mapping,
elevation level, and 3D mapping are some examples in which several images are stitched
together. Moreover, the image analysis process can be repeated periodically what will
result in the visual representation of the evolution of a particular area or item over time.
Nevertheless, image do not supply accurate measurements to what they represent. To add
accurate quantitative data to images, these can be combined with LiDAR (Fei et al. 2008)
or with sequenced BIM models (Skibniewsk 2014).
CHRONO-ANALYSIS
Chrono-analysis is the assessment of footages to evaluate production. The footages are
captured by cameras positioned around the shop floor to record an activity done by the
worker(s). The time spent by the worker(s) on each task of the activity can be measured
by watching the recordings. The tasks are classified into three major categories: value-
adding, non-value-adding but necessary (Muda – Level I), and non-value-adding and
unnecessary (Muda – Level II). One of the mantras of Lean is “eliminate Muda.”
Accordingly, first, the analyst will identify each task that composes the activity. For
instance, the footage of the activity contains the recordings of the set-up, the core task,
breaks and the cleaning. Next, the analyst chronographs each task. Then, the analyst plan
on how to eliminate or at least minimize the time spent on non-value-adding tasks (set-up,
breaks, and cleaning). Then, the analyst plan on how to eliminate or at least minimize the
time spent on non-value-adding tasks (set-up, breaks, and cleaning). Later, the analyst
implements the plan and potential solutions. The analyst's records new footage of the

activity execution and tracks the time spend on the tasks. After a comparison of the times
to the original results, the analyst updates the activity standard with the solutions that
resulted in improvement. Chrono-analysis can be seen as lean focused, more detailed, and
evidential implementation of time and motion analysis. The caveats: chrono-analysis is
usually a laborious process conducted eventually rather than continuously; the benefits
for tasks with a low level of repetitiveness are minuscule.
PRODUCTION THEORY
The traditional theory about fundamental mechanisms of production in repetitive
processes in construction is at an embryonic stage and does not yet fully establish the
foundations of a production model. The traditional and convenient approach to project-
driven production in construction is to rely on linear steady state models. By considering
the transient state, Productivity Function produces models that are more accurate in
describing the processes dynamics than the steady state approaches (Antunes et al. 2017).
The Productivity Function provides a mathematical foundation to develop algebraic for
the calculations of cycle times (average, best- and worst-cases), throughput at capacity
(Antunes et al. 2018), and the influence of the transient state time in the production
variability (Antunes et al. 2016).
Productivity Function has been applied in feedback loop control yielding a controlling
approach [Productivity Function Predictive Control (PFPC)] that can achieve high
performances even when processes operate closer to capacity (Antunes 2017). Moreover,
this performance enhancement is higher when PFPC is applied to processes in a parade-
of-trades (Tommelein 1998). The PFPC shown to be a robust approach to plan, control,
and optimize production and supply chain in construction with direct implications to
management practices such as takt time. A benefit of PFPC is its focus on minimizing the
variances of output to the set point or plan. The PFMPC can operate satisfactorily even
without an accurate model (Antunes 2017). In practice, the use of adaptive PFPC
(APFPC) can be useful. This adaptive version estimates a Productivity Function
cyclically within a period; thus, the control relies on a model that is accurate to the
current time frame. Therefore, if the production system evolves (which is the goal of
continuous improvement) that makes the model obsolete, APFPC can relearn the process
and estimate a new model automatically.
Although the Productivity Function can describe a variety of systems (including
multi-variables systems), a structure that can embrace nonlinear and/or time-variant
systems is required; and respectively, the introduction of linear time-varying space-state
models which can also describe nonlinear systems. Nevertheless, the evaluation of these
function from the data is based on the back-propagation algorithm (Antunes 2017), which
is a machine learning tool.
MACHINE LEARNING
Machine learning is the term used to describe a field in computer science where the
machine is trained on how to do a task instead of being programmed. Thus, by being
trained (or training itself) the machine can develop its own way of how to execute the
task (Silver et al. 2017). The training can be either assisted or unassisted. Assisted

training means that the inputs and outputs are provided to the machine that makes sense
of the conditions to determine the output. For unassisted training, only the inputs are
available. That entails enormous flexibility to machine learning and its applicability. As
such, machine learning can mix a variety of input sources (features) to determine or
classify the output, being capable of performing simple (such as an and operation) to
complex tasks. For instance, it can evaluate labor processes as numerical values to
estimate a non-linear productivity function (Antunes 2017), or identify and track different
elements at once in a video feed (Gordon et al. 2017).
FRAMEWORK
The top-down and bottom-up approaches interact joining theory and practice in
continuous improvement loop. This suggestion stands on two tenets: observer effect and
Genchi Genbutsu. In physics, the term observer effect (Bianchi 2013) defines the
influence of the observation act to the event. It means that by observing an event, the
observer may alter the event, and consequently modify the observation. This effect is also
known in the human sciences, where subjects have their behavior affected by being
observed. In this sense, the awareness of being observed may modify the production
system and its model. Thus, production is constantly observed, and the information is
used to modify production. Genchi Genbutsu, a principle of the Toyota Production
System, which means ‘go to the source and get the facts to make the right decision.’ In
this approach, instead of asking for information updates the progress status is obtained in
real time from positioned sensors or upon inspection from the drones. Next, the machine
learning engine will merge the information (LiDAR, images, sensors) with BIM to
identify the product progress and deviations from the specifications (similarly as in the
SCADA). In parallel, the production information (progress and workers information) is
checked against the production theory and models to evaluate productivity, forecast
conclusion dates and assess corrective actions (as in APFPC). These two combined and
jointly with the project plan are then presented to the ‘control room.’ Therefore, the
‘control room’ can rely on accurate information in the decision-making process, which
results in a data-driven continuous improvement loop (Figure 1).
For instance, if a fixed camera detects that a disposal bin is being filled at a certain
rate the replacement of that bin can be ordered from the control room without the
worker’s requisition (that means eliminating the requisition task (Muda Level I), the work
stoppage (Muda Level II) by waiting that the bin replacement or having to replace it
(setup/cleaning, i.e., Muda Level I). And as Muda decreases Mura also decreases
(Antunes et al. 2016). Similar reasoning works with suppliers. For example, if the casing
were not cemented in place, the suppliers can be notified to avoid bringing more to the
site. This integration with the supplier may avoid Muda (Level I) in one or more
situations: inventory - use extra space to store more that its needed; waiting - if the trucks
need to wait around the site; motion - case the truck needs to go back. Because
information is compiled in the SCADA and centralized in the ‘control room’ it can be
accessed and shared with ease, such as in a library.

Figure 1 Theoretical framework for an information integration system for construction
CONCLUSIONS
Cross-functional cooperation in construction is low mostly because the parts have no
information about what is happening outside their area. The same can be said about
suppliers. The establishment of the ‘control room’ centralizes information from the plan,
labor, process, and production. Moreover, once the control-room has information about
the progress of the current and next activities on-site, it will be able to coordinate cross-
functional activities and supply chain.
Building information security and maintenance may use the product legacy
information gathered in construction eliminating redundant work by analyzing the
building. This work has already been done during construction (reducing over-
processing). There is a compiled log of who did what, when and how for every part of the
building including divergences between the original design and every change and defect
occurred during construction. There can be extensive details of how the process has been
done (and evolved). The production knowledge has further benefits. Especially, due to
the network effect. The network effect adds value to this framework with use and
adoption. This means that data can be generalized to a broader audience with more
information such as season of the year, weather condition, geo-localization, altitude,

winds, local culture, diversity, or any other feature. Hence, future endeavors will establish
the production base-line using historical evidence rather than the usual labor/time
relationship.
Using the chrono-analysis continuous assessment jointly with the data (production
progress and workers effort) from previous projects informal processes tends to be
eliminated. Better processes are developed and standardized. More accurate historical
information is persistent and can be generalized to different projects enabling comparison
and continuous improvement methodologies from project to project. New builders will be
trained in the benchmark process instead of the “I have been done this for the last x
years” (and repeating the same mistakes over and over) approach. As such, the
conservative company culture, lack of innovation and delayed adoption will be addressed
by the marked. Companies will quantitatively assess and qualify the performance of
contractors in previous projects. In an intensive third-party contracting industry such as
construction, low productivity companies that often make mistakes are costly, and
consequently, put at the end of the supplier's list or dismissed. Construction needs an
increase in the number of builders, but it really needs builders with better performance.
A more automated construction industry should experience a set of benefits, such as
better decision-making processes, increase information flow, and increase productivity.
These benefits have a collateral impact on the whole society. More productivity means
that more projects can be done using fewer resources. Accordingly, more infrastructure
can be built and maintained. It can increase the affordability of the housing prices.
REFERENCES
Almeida, P. R. de, and Solas, M. Z. (2016). Shaping the Future of Construction: A
Breakthrough in Mindset and Technology. Davos, Switzerland, 1–64.
Antunes, R. (2017). “Dynamics of project-driven systems: A production model for repetitive
processes in construction.” PhD Thesis, The University of Auckland.
Antunes, R., and Gonzalez, V. (2015). “A Production Model for Construction: A Theoretical
Framework.” Buildings, 5(1), 209–228.
Antunes, R., González, V. A., Walsh, K., and Rojas, O. (2017). “Dynamics of Project-Driven
Production Systems in Construction: Productivity Function.” Journal of Computing in
Civil Engineering, 31(5), 17.
Antunes, R., González, V. A., Walsh, K., Rojas, O., O’Sullivan, M., and Odeh, I. (2018).
“Benchmarking Project-Driven Production in Construction Using Productivity Function:
Capacity and Cycle Time.” Journal of Construction Engineering and Management,
144(3), 04017118.
Antunes, R., González, V., and Walsh, K. (2016). “Quicker reaction, lower variability: The
effect of transient time in flow variability of project-driven production.” 24th Annual
Conference of the International Group for Lean Construction, Boston, MA, sect.1 pp.
73–82.
Autodesk. (2002). “Building Information Modelling.”
<http://laiserin.com/features/bim/autodesk_bim.pdf> (Feb. 15, 2018).
Bianchi, M. S. de. (2013). “The Observer Effect.” Foundations of Science, 18(2), 213–243.

Eastman, C. M. (1969). “Cognitive processes and ill-defined problems: A case study from
design.” Proceedings of the International Joint Conference on Artificial Intelligence:
IJCAI, 69, 669–690.
Eastman, C. M. (1973). “Automated space planning.” Artificial Intelligence, 4(1), 41–64.
Engelbart, D. (1962). Augmenting Human Intellect: A Conceptual Framework. Stanford
Research Institute, Melo Park, CA, 1–29.
Fei, D., HuMinjie, and HaiYan, G. (2008). “Automatic Registration Between LiDAR and
Digital Images.” The International Archives of the Photogrammetry, Remote Sensing and
Spatial Information Sciences Vol. XXXVII. Part B1, Beijing, China, 487–490.
Gordon, D., Kembhavi, A., Rastegari, M., Redmon, J., Fox, D., and Farhadi, A. (2017). IQA:
Visual Question Answering in Interactive Environments.
Hayes, R. H., and Wheelwright, S. C. (1979). “Link manufacturing process and product life
cycles.” Harvard Business Review, 57(1), 133–140.
Kopetz, H. (2011). Real-time systems design principles for distributed embedded applications.
Springer, New York, NY.
McGraw Hill Construction. (2012). The Business Value of BIM in North America Multi-Year
Trend Analysis and User Ratings (2007–2012). McGraw-Hill Construction, Bedford, MA.
National BIM. (2017). National BIM Report 2017. RIBA Enterprises Ltd, Newcastle, 22.
National Society of Professional Engineers. (2014). “Construction Productivity in Decline.”
The Magazine For Professional Engineers, 13.
Nederveen, G. A. van, and Tolman, F. P. (1992). “Modelling multiple views on buildings.”
Automation in Construction, 1(3), 215–224.
Rossini, F. L., Novembri, G., and Fioravanti, A. (2017). “BIM and Agent-Based Model
Integration for Construction Management Optimization.” 25th Annual Conference of the
International Group for Lean Construction, Heraklion, Greece, 111–118.
Silver, D., Schrittwieser, J., Simonyan, K., Antonoglou, I., Huang, A., Guez, A., Hubert, T.,
Baker, L., Lai, M., Bolton, A., Chen, Y., Lillicrap, T., Hui, F., Sifre, L., van den
Driessche, G., Graepel, T., and Hassabis, D. (2017). “Mastering the game of Go without
human knowledge.” Nature, 550(7676), 354–359.
Skibniewsk, M. J. (2014). “Construction Project Monitoring with Site Photographs and 4D
Project Models.” Organization, technology and management in construction: An
international journal, 6(3).
Tommelein, I. D. (1998). “Pull-Driven Scheduling for Pipe-Spool Installation: Simulation of
Lean Construction Technique.” Journal of Construction Engineering and Management,
124(4), 279–288.
Wang, J., Sun, W., Shou, W., Wang, X., Wu, C., Chong, H. Y., Liu, Y., and Sun, C. (2015).
“Integrating BIM and LiDAR for Real-Time Construction Quality Control.” Journal of
Intelligent and Robotic Systems: Theory and Applications, 79(3–4), 417–432.
Wesolowski, Z. P. (1978). “Polish Contribution To Management Science.” Academy of
Management Journal, 12–16.

Envision of an Integrated Information System for Projectdriven Production in Construction

More Related Content

What's hot (20)

Similar to Envision of an Integrated Information System for Projectdriven Production in Construction (20)

More from Ricardo Magno Antunes (13)

Recently uploaded (20)

Envision of an Integrated Information System for Projectdriven Production in Construction