Session 1121

Impact of 4D Visualization on the Cognitive Process

of Detecting the Logical Errors in the Construction Schedule

Julian H. Kang* and Byeong-Cheol Lho**

*Texas A&M University / **Sangji University

Abstract

In order to understand the construction schedule, students have to read drawings, visualize the
structures in mind, and link these structures with the schedule information depicted in the bar
chart or CPM network. Since this is not a simple process, students studying construction
scheduling sometimes hardly integrate the bar chart and the two-dimensional (2D) drawings to
visualize the construction sequence in their mind. Four-dimensional (4D) visualization, which
shows visually the sequence of building the structures over time using three-dimensional (3D)
computer graphics, is expected to be one of the innovative methods that improve our cognitive
process of understanding the construction schedule.
This paper introduces an experiment to test whether the students would better understand the
construction schedule if 4D visualization is provided. For this experiment, students studying
construction management at Texas A&M University were recruited and then asked to detect
logical errors in the construction schedule using different levels of graphical representation. One
group conducted the experiment using 4D visualization of the construction sequence, and the
other group used 3D model of the constructed structure. For the experiment, an Internet
instrument was developed to provide graphical representation of the construction schedule and
measure the elapsed time for the students to detect logical errors in the schedule. The experiment
showed that students who used 4D visualization detected more logic errors within less time than
those who used 3D visualization.

I. Introduction

Construction planning requires reading the drawings, visualizing the constructed structures in
mind, breaking the structures into identifiable components, and building a logical network among
these components. Once the duration required for constructing each component is estimated, the
computer application can identify the critical path in the network and calculate the total duration
of the project using Critical Path Method (CPM). The construction schedule is then conceptually
illustrated using the bar chart. The conceptual expression of the schedule has been considered
effective for illustrating the entire construction schedule. However, it may take many years to
develop a skill for understanding the complicated construction sequence and detecting any logical
errors hidden in the construction schedule if the construction schedule is depicted using the bar
chart. Undetected logical errors in the schedule could delay the entire construction project.
Project engineers try to proactively detect any logical errors hidden in the construction schedule
Page 8.652.1

Proceedings of the 2003 American Society for Engineering Education Annual Conference & Exposition
Copyright © 2003, American Society for Engineering Education
to finish the project on time.

The cognitive process of detecting the logical errors in the construction schedule is similar to the
process of developing the construction schedule. One must read the drawings, visualize the
structure in mind, break the structure into identifiable components, and then link these
components with the construction schedule illustrated by the bar chart. This may be an easy
process for the experienced constructors. However, the owners or the end users may have
difficulties in integrating the bar chart and the 2D drawings to understand the construction
schedule although they may have better knowledge of utilizing the constructed facility. Therefore
it may not be reasonable to expect that they could be part of the process of making proactive
decisions in construction planning. They may need comprehensive tools to understand the
increasing complexity of the constructed product.

It is known that visualization may improve the human’s cognitive process of understanding the
construction schedule. Murgio9 insisted that human beings obtain 83% of their knowledge from
visual observation. Johnson-Laird et al.5 empirically verified that the use of realistic materials
improved performance in a deceptive reasoning problem. Pressley12 asserted that imposed pictures
are almost always learned better than words.

Recent research tried to show the construction schedule visually using three-dimensional (3D)
computer model1, 13, 15. Bechtel Corporation integrated 3D Computer Aided Design (CAD) models
with scheduling packages to simulate the construction operations7, 13. Bechtel later developed 4D-
Planner, a graphical simulation tool, that helps project managers, construction planners, and field
engineers plan and manage their projects effectively15. The Center for Integrated Facility
Engineering (CIFE) at Stanford University also integrated 3D CAD objects with the construction
schedule to show the construction sequence visually in 3D environment1. The CIFE applied 4D
CAD to the San Mateo County Rehabilitation Center campus expansion project2 and a new Walt
Disney Concert Hall designed by Frank O. Gehry4, and demonstrated that 4D visualization of the
construction sequence helped project participants proactively manage their projects.

More recent research attempted a new way of distributing 4D visualization using the Internet
technology. Kang6 developed a Web-based 4D visualization application using Java 3D
technology. With this software, end users can update the construction schedule over the Internet
and the 4D visualization model of the updated construction schedule can be displayed immediately
on the Web browser.

Many construction firms, however, still hesitate to utilize 4D CAD for their project due to the
initial investment of creating a 3D CAD model and lack of confidence that they could make profit
from utilizing 4D CAD. It is necessary, therefore, to provide empirical evidence for the impact of
4D visualization in construction planning. This paper presents an experiment to measure the
impact of 4D visualization on human’s cognitive process of understanding the construction
schedule.
Page 8.652.2

Proceedings of the 2003 American Society for Engineering Education Annual Conference & Exposition
Copyright © 2003, American Society for Engineering Education
II. Review of 4D visualization for construction planning

Two-Dimensional (2D) drawings have been used in the architecture, engineering and construction
(AEC) industry to describe a designer’s idea for the constructors. However, in order to visualize
the shape of the structure in a three-dimensional (3D) world, one has to read several drawings to
comprehend the spatial and volumetric aspects of the objects. This process requires a significant
amount of education in the conventions of drawing and practical experience. Developing a
construction schedule is even harder because one must build a structure step-by-step in mind after
visualizing it in a 3D world. 3D objects can be best described by 3D visualization14. Architects
therefore have extensively used a 3D miniature model to show their design to the customers and
get feedback from the customers more effectively. As 3D CAD technology became available in
the AEC industry, architects learned that they can build and modify a 3D model quickly using 3D
CAD, and distribute it to the project participants easily. They also noticed that they can combine
additional engineering information with the 3D CAD model. Engineering, procurement and
construction (EPC) firms were leaders in integrating 3D CAD with engineering information.
Bechtel Corporation’s 3D modeling system for plant design (3DM), and Stone & Webster’s
Construction Management Display System (CPMANDS), Fluor Daniel’s CALMA Plant Design
System (PDS), and Black & Veatch’s POWERTRAK are some of examples of integrating 3D
CAD with design information. These systems were used for reviewing constructibility, checking
interferences, and taking off the bill of materials7.

Bechtel Corporation expanded the use of their 3DM to simulate the construction operations and
assembly sequence13. The system includes construction equipment models and allows the planner
to include temporary structures or facilities to accurately simulate the construction environment.
It also provides a dynamic interference checking function to validate proposed construction flows
and equipment selections7. Bechtel Corporation also developed 4D-Planner that imports the 3D
CAD model and the construction schedule from the commercial applications, such as
MicroStation, Plant Design System (PDS), AutoCAD, and Primavea Project Planner, in order to
merge them into a simulation file that can be reviewed interactively. The 4D-Planner is reported to
help project managers, construction planners, and field engineers plan and manage their projects
effectively15.

The Center for Integrated Facility Engineering (CIFE) at Stanford University connected 3D CAD
objects with the construction schedule to develop a graphical representation that shows the
construction sequence1. The CIFE and Dillingham Construction jointly demonstrated in the San
Mateo County Rehabilitation Center campus expansion project that 4D CAD helped people
understand a construction schedule intuitively2. It was found that even doctors and nurses were
able to understand the impact of construction plans on their department, office, and daily
operations. The CIFE also utilized 4D CAD for the new Walt Disney Concert Hall construction
project, which was designed by Frank O. Gehry. Using 4D CAD, the project team was able to
visualize several what-if scenarios to detect conflicts before the project began, and effectively
explain why certain decisions were made to the board members who were unfamiliar with the
construction process4, 11.
Page 8.652.3

Proceedings of the 2003 American Society for Engineering Education Annual Conference & Exposition
Copyright © 2003, American Society for Engineering Education
Balfour Technologies LLC developed fourDviz which was used by the National Aviation and
Transportation Center, Brookhaven, NY, and Frederic R. Harris Inc., Boston, Mass., for several
projects, including the LaGuardia Airport Rail Access project and Boston's Logan Airport
Modernization Program. Balfour Technologies LLC also plans to develop a 4D browser that will
allows users to access a 4D portal via the Web or corporate intranet4.

A design and construction of Helsinki University of Technology Auditorium Hall 600 (HUT-600)
in Finland used commercially available state-of-the-art analytical and visualization tools to
optimize the design, construction, and operation of the facility during early project phases. The
project team was able to shorten the time for design iteration and develop a reliable budget for
effective cost control. Visualization tools such as Virtual Reality-Experimental Virtual
Environment (VR-EVE) fostered early communication among the end users, which eventually
helped the project team capture valuable inputs from the clients and effectively utilize them for the
design work. The HUT-600 demonstrated that 4D visualization could expedite the conventional
design practices and promote the life-cycle consideration in the construction project3.

However, some limitations of 4D CAD were revealed while using 4D CAD in the construction
project, such as the necessity of a large initial investment of time and effort to create a 3D model,
and limitations in updating the 4D CAD model8. Kang6 developed a prototype of Web-based 4D
visualization software using Java 3D and Java JDBC technology in order to overcome some of
the limitations of current 4D CAD. This software enables end-users to modify the construction
schedule over the Internet and confirm the revised construction sequence visually on a Web
browser in 3D environment.

III. Experiment design

The impact of 4D visualization was mainly investigated through comparison of the level of
effectiveness between two different graphic representations in helping the experiment participants
detect logical errors hidden in the construction schedule. For this investigation, graduate students
studying construction management at Texas A&M University were recruited. The students were
randomly divided into two groups, group A and group B, which used different graphic
representations for illustrating the construction schedule. The students were then asked to detect
logical errors in the sequence of building wooden toy block towers with the assumptions that 1)
no adhesive is used to put two blocks together and 2) the wooden blocks should be placed one-by-
one using only one hand. The constructed wooden towers used in the experiment are illustrated in
Figure 1. The logical errors were created by making some blocks unsupported in the building
process.
Page 8.652.4

Proceedings of the 2003 American Society for Engineering Education Annual Conference & Exposition
Copyright © 2003, American Society for Engineering Education
 Figure 1: Wooden toy tower model used in task 1 (left) and task 2 (right)

In order to investigate the impact of 4D visualization, each group was provided different levels of
resources. One group was provided 1) 2D drawings that illustrate the plans and elevations of the
wooden towers; 2) a 3D computer model of the tower that enables the experiment participants
navigate around the tower; 3) wooden block numbers that represent the sequence of stacking up
the wooden blocks to build a tower. The other group was provided 1) the same 2D drawings, and
2) the same wooden block numbers. However, this group was provided a 4D computer model so
that the experiment participants can navigate not only around the tower model, but also over time
to observe the construction sequence visually. Figure 2 shows the sequence of building the tower
using the 4D computer model.

The experiment was implemented on the Web browser. All instructions and necessary graphic
representations were provided on the Web browser. The experiment participants were asked to
provide the number of logical errors they detected via the Web page, and their answers were
collected in the database located in the Web server along with a time stamp. Figure 3 shows some
of the Web-based instruments used in the experiment. The collected answers were then analyzed
to identify the following data:

• Number of logical errors detected

• Average elapsed time to detect one logical error

The result of the experiment in this research can be easily biased due to huge individual
differences between the participants10. Subject A that uses a 3D computer model could finish the
given task faster than subject B that uses a 4D computer model simply because subject A is more
experienced in using the given resources than subject B. Therefore, the experiment was designed
to have every participant use both graphic representations for problem solving. The experiment
participants were asked to answer two questions. If a certain participant uses a 4D computer
model in the first question, this subject is then provided a 3D computer model in the second
question. If a subject starts solving the question with a 3D computer model, the subject uses a 4D
Page 8.652.5

Proceedings of the 2003 American Society for Engineering Education Annual Conference & Exposition
Copyright © 2003, American Society for Engineering Education
model in the second question.

Figure 2: 4D model of wooden toy tower

Figure 3: Web-based Instrument with 4D model (left) and 3D model (right)

Page 8.652.6

Proceedings of the 2003 American Society for Engineering Education Annual Conference & Exposition
Copyright © 2003, American Society for Engineering Education
Table 1 shows the type of graphic representations provided to each group. It is expected that
participants who are particularly talented or untalented should not skew the results, as they affect
both sets of statistics equally. For training purposes, one exercise problem was given to the
experiment participants prior to starting the actual tasks.

Table 1: Type of graphic representations used in the experiment

Group Task 1 Task 2

Group A 4D model 3D model
Group B 3D model 4D model

After finishing the given tasks, the experiment participants were asked to provide their opinion
about using 4D visualization for conducting the experiment. The following questions were asked
in the exit survey:

1. What's your gender? (Male/Female)

2. How old are you? (21-25/26-30/31-35/36-40)
3. What's your classification? (Master/PhD)
4. Have you ever taken any construction scheduling class? (Yes/No)
5. Have you ever built a schedule for the real project? (Yes/No)
6. Have you ever taken any 3D CAD modeling class? (Yes/No)
7. When the 3D model was available during the experiment, how easy was it to understand
the sequence of building the toy tower? (1: Very difficult – 10: Very easy)
8. When the 4D model was available during the experiment, how easy was it to understand
the sequence of building the toy tower? (1: Very difficult – 10: Very easy)

IV. Experiment results

A total of 20 students participated in the experiment. Among 20 students 16 students finished the
experiment as per the instruction. Table 2 shows the number of students in each group and the
type of resources they used in the experiment. For the analysis, the measurements are reorganized
by the type of graphic representation that the experiment participants used: a 3D group and 4D
group. Table 3 shows the number of measurements in each task based on the new arrangement.

Table 2: Number of subjects in each group and type of graphic representation used

Group No. of Graphic Representation

Subjects
Task 1 Task 2
Group A 8 4D model 3D model
Group B 8 3D model 4D model
Page 8.652.7

Proceedings of the 2003 American Society for Engineering Education Annual Conference & Exposition
Copyright © 2003, American Society for Engineering Education
 Table 3: Number of measurements in each task

Group Number of measurements

Task 1 Task 2
3D Group 8 8
4D Group 8 8

In the experiment, all answers that the participants provided were saved in the database of the
Web server with a timestamp. The following raw experiment measurements were collected in each
task:

• Time that the subject started the task

• Time that the subject finished the task
• Number of logical errors that the subject detected in the task

The raw measurements were used to generate the following data in each task:

• Number of logical errors detected

(= Number of actual logical errors that the subject detected in the task)
• Average elapsed time to detect one logical error
(= (Time that the subject finished the task – Time that the subjct started the task) /
Number of logical errors that the subject detected in the task)

The results show that the students detected all logical errors when they used the 4D model. When
the students used the 3D model, four students either did not detect all logical errors or made a
mistake in detecting the logical errors. In general, those who used the 4D model took less time to
detect the logical errors than those who used the 3D model. As shown in Figure 4, the students
spent 1 minute 5 seconds to detect one logical error average when they used the 4D model in the
first task. The students spent 2 minutes 36 seconds in average when they used the 3D model in the
same task. In the second test, the average elapsed time spent to detect one logical error was
significantly reduced regardless of graphic representations. It is asssumed that elapsed time was
reduced because of learning effect. In the second question, the 3D group and 4D group marked 1
minute 9 seconds and 0 minute 29 seconds respectively to detect one logical error.

Although the number of subjects is not large enough to conduct a statistical test to show if there
is a significant difference between two groups, it is distinctive that the 4D group detected the
logical errors faster than the 3D group in all tasks. Since this experiment is currently being
conducted with more subjects, the statistical evidence is expected to be provided in the near
future.

For the question of the exit survey, “When the 3D model was available during the experiment,
how easy was it to understand the sequence of building the toy tower? (1: Very difficult – 10:
Very easy)”, the average point of the participants marked 6.40 out of 10.
Page 8.652.8

Proceedings of the 2003 American Society for Engineering Education Annual Conference & Exposition
Copyright © 2003, American Society for Engineering Education
 0:02:53

0:02:36
0:02:18
0:02:01

Elapsed Time
0:01:44

0:01:26
0:01:09
0:00:52
0:00:35
0:00:17
0:00:00
1 2

3D 0:02:36 0:01:09
4D 0:01:05 0:00:29
Task

Figure 4: Average elapsed time to detect one logical error

For the question, “When the 4D model was available during the experiment, how easy was it to
understand the sequence of building the toy tower? (1: Very difficult – 10: Very easy)”, the
average point marked 9.27 out of 10. This result clearly shows that the experiment participants
felt more comfortable when they used the 4D model for detecting the logical errors in the given
tasks.

V. Concluding remarks and future work

The results of the experiment demonstrated a quantitative evidence that 4D visualization helped
the students understand the construction schedule. Repeating the experiment with more subjects is
expected to produce a strong evidence that would convince construction professionals to consider
the utilization of 4D visualization for construction planning. The experiment also will be
conducted using different population in order to secure the realibility and validity of the test and
indicate sensitivity of the instrument. The impact of 4D visualization on learning is another
provocative area of research. One may speculate that 4D visualization would be more effective for
teaching principles of project planning and scheduling because it is less abstract than CPM
networks or bar charts.

Bibliography

1. Collier, E. and Fischer, M., “Four-Dimensional Modeling in Design and Construction,” Technical Report, No.
101, Center for Integrated Facility Engineering, Stanford University, Stanford, 1995.
2. Collier, E. and Fischer, M., “Visual-Based Scheduling: 4D Modeling on the San Mateo County Health
Center,” Computing in Civil Engineering: Proceedings of the Third Congress Held in Conjunction with A/E/C
Systems '96, Ed. Jorge Vanegas and Paul Chinowsky, American Society of Civil Engineers, New York, 1996,
pp. 800-805.
3. Fischer, M. and Kam, C., “PM4D Final Report,” CIFE Technical Report, Number 143, Center for Integrated
Page 8.652.9

Proceedings of the 2003 American Society for Engineering Education Annual Conference & Exposition
Copyright © 2003, American Society for Engineering Education
 Facility Engineering, Stanford University, Stanford, 2002.
4. Goldstein, H., “4D: Science Fiction or Virtually Reality?” Construction.com, (http://www.construction.com/
NewsCenter/it/features/01-20010416.asp), 2001.
5. Johnson-Laird, P. N., Legrenzi, P., and Legrenzi, M. S., “Reasoning and a Sense of Reality,” The British
Journal of Psychology, Vol. 63(3), University Press, Cambridge, 1972, pp.395-400.
6. Kang, H., “Web-based 4D Visualization for Construction Schedulin g,” Ph.D. Dissertation, Texas A&M
University, College Station, TX, 2001.
7. Mahoney, J.J., Tatum, C.B., and Kishi, K., “Construction Site Applications of CAD,” Technical Report,
Number 36, Center for Integrated Facility Engineering, Stanford University, Stanford, 1990.
8. McKinney, K., Kim, J., Fischer, M., and Howard, C., “Interactive 4D-CAD,” Computing in Civil
Engineering: Proceedings of the Third Congress Held in Conjunction with A/E/C Systems '9 6, Ed. Jorge
Vanegas and Paul Chinowsky, American Society of Civil Engineers, New York, 1996, pp.383-389.
9. Murgio, M.P., Communications Graphics, Van Nostrand Reinhold, New York, 1969.
10. Nielsen, J., Usability Engineering, Academic Press, Inc., San Diego, 1993.
11. Post, N.M., “Movie of Job that Defies Description Is Worth More Than A Million Words,” ENR: Engineering
News-Record, Vol. 248 No. 13, The McGraw-Hill Companies, Inc., 2002.
12. Pressley, M., “Imagery and Children’s Learning: Putting the Picture in Developmental Perspective,” Review of
Educational Research, Vol. 47(4), 1977, pp.585-622.
13. Simons, K.L., Thornberry, H.L., and Wickard, D.A., “Simulation System for Construction Planning and
Scheduling,” Joint ASME/IEEE Power Generation Conference, Philadelphia, 1988.
14. Tufte, E.R., Envisioning Information, Graphics Press, Cheshire, 1990.
15. Williams, M., “Graphical Simulation for Project Planning: 4D-PlannerTM,” Computing in Civil Engineering:
Proceedings of the Third Congress Held in Conjunction with A/E/C Systems '96, Ed. Jorge Vanegas and Paul
Chinowsky, American Society of Civil Engineers, New York, 1996, pp.404-409.

JULIAN H. KANG
Julian H. Kang is an assistant professor of the Department of Construction Science at Texas A&M University. His
research interests include the utilization of the Internet technology and mobile devices for construction planning
and project information management. For collaborative construction planning, he is interested in using Java 3D
and XML for visualizing the construction sequence on the Web browser.

BYUNG-CHEOL LHO

Byeong Cheol Lho is an associate professor of the Department of Civil Engineering, and a director of the Disaster
Prevention Institute at Sangji University. He has investigated the utilization of the Internet technology for structure
design and maintenance. He is currently developing a Web-based computer application for the design and
maintenance of simple concrete structures.

Page 8.652.10

Proceedings of the 2003 American Society for Engineering Education Annual Conference & Exposition
Copyright © 2003, American Society for Engineering Education