www.elsevier.com/locate/compind
Available online at www.sciencedirect.com
(2008) 477–488
Computers in Industry 59A versatile virtual prototyping system for rapid product development
S.H. Choi *, H.H. Cheung
Department of Industrial and Manufacturing Systems Engineering, The University of Hong Kong, Pokfulam Road, Hong Kong
Received 25 January 2007; received in revised form 22 June 2007; accepted 17 December 2007
Available online 1 February 2008
Abstract
This paper presents a versatile virtual prototyping (VP) system for digital fabrication of multi-material prototypes to facilitate rapid product
development. The VP system comprises a suite of software packages for multi-material layered manufacturing (MMLM) processes, including
multi-toolpath planning, build-time estimation and accuracy analysis, integrated with semi-immersive desktop-based and full-immersive CAVE-
based virtual reality (VR) technology. Such versatility makes the VP system adaptable to suit specific cost and functionality requirements of
various applications.
The desktop-based VR system creates a semi-immersive environment for stereoscopic visualisation and quality analysis of a product design. It
is relatively cost-effective and easy to operate, but its users may be distracted by environmental disturbances that could possibly diminish their
efficiency of product design evaluation and improvement. To alleviate disturbance problems, the CAVE-based VR system provides an enclosed
room-like environment that blocks out most disturbances, making it possible for a design team to fully concentrate and collaborate on their product
design work.
The VP system enhances collaboration and communication of a design team working on product development. It provides simulation
techniques to analyse and improve the design of a product and its fabrication processes. Through simulations, assessment and modification of a
product design can be iterated without much worry about the manufacturing and material costs of prototypes. Hence, key factors such as product
shape, manufacturability, and durability that affect the profitability of manufactured products are optimised quickly. Moreover, the resulting
product design can be sent via the Internet to customers for comments or marketing purposes. The VP system therefore facilitates advanced product
design and helps reduce development time and cost considerably.
# 2007 Elsevier B.V. All rights reserved.
Keywords: Virtual prototyping; Virtual reality; Immersive visualisation; Product design evaluation; Multi-material layered manufacturing
1. Introduction
Mounting pressure of intensifying market globalization and
competition has been driving manufacturing industries to
compete on incessant reduction in lead-time and cost of product
development while assuring high quality and wide varieties.
However, conventional manufacturing processes are no
longer sufficient to speed up validation of product design and
development processes to meet ever-increasing diversities of
customer demands, stringent cost control, and complexity of
new products.
Indeed, the significance of rapid product development or
rapid manufacturing has been recognised in recent years. For
this, many researchers have worked on developing various
* Corresponding author.
E-mail address: [email protected] (S.H. Choi).
0166-3615/$ – see front matter # 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.compind.2007.12.003
technologies, which can be roughly categorised into three
areas:
(i) L
ayer manufacturing (LM) technology for physicalfabrication of product prototypes, rapid tooling, and direct
manufacture of components;
(ii) H
eterogeneous object modelling schemes and multi-material toolpath generation algorithms for design and
subsequent fabrication of composite and functionally
graded objects, such as bio-degradable scaffolds;
(iii) V
irtual prototyping (VP) and virtual manufacturing (VM)simulation techniques for digital fabrication of prototypes,
validation and optimisation of product designs, and
evaluation of product assemblability and producability.
Among these technologies, virtual simulation is regarded as an
important technological advancement for product develop-
ment, and it has been successfully used in ship-building and car
S.H. Choi, H.H. Cheung / Computers in Industry 59 (2008) 477–488478
industries [1,2]. It is a process of using virtual prototypes and
simulation techniques, often in a virtual reality (VR) system
with innovative input and stereoscopic output, to evaluate and
improve a product design and to validate its planning and
manufacturing processes [3–8]. Through simulations, key
factors such as the shape and the manufacturability of a product
may be optimised without committing much to prototypes and
tooling. Indeed, virtual simulation reduces the need for physical
prototypes and hence minimizes tooling cost and material
waste, and it allows manufacturers to ‘‘get it right the first time’’
and helps them deliver quality products to market on time and
within budget.
However, the current virtual simulation technique, which
often adopts either semi-immersive or full-immersive VR, is
not without limitations, particularly with respect to the sense of
immersion.
Bochenek and Ragusa [6] pointed out that it is important to
appropriately select a VR system for product design and
development processes. They investigated the use of four
commercial VR display systems and found that the sense of
immersion plays an important role in improving the design
review practices, and that a higher sense of immersion
facilitates better improvement.
In general, semi-immersive VR systems (single-screen or
desktop-based) are relatively easy to use, affordable, and of
good resolution, though their users tend to be susceptible to
environmental distractions. On the other hand, full-immersive
VR systems (multi-screen or CAVE-based) can generate a
relatively higher sense of immersion that facilitates user
interaction and collaboration, but they are generally more
expensive, of less resolution and poor portability, and needs
special space requirements [4,9,10]. Hence, it is worthwhile to
combine the good features of both semi- and full-immersive VR
to enhance the versatility and effectiveness of virtual simulation
at affordable cost.
This paper therefore proposes a versatile virtual prototyping
system for evaluation of product designs and digital fabrication
of multi-material prototypes either in a semi-immersive
environment or in a full-immersive, disturbance-free environ-
ment to facilitate rapid product development. The versatility of
choosing between semi- and full-immersive VR environments
makes the VP system adaptable to suit the cost and
functionality requirements of various applications.
The VP system integrates semi- and full-immersive VR
with multi-material layered manufacturing (MMLM) tech-
nologies. It comprises mainly a suite of software packages for
simulation of MMLM processes, including multi-toolpath
planning, build-time estimation, accuracy analysis, a desktop-
based VR system, and a CAVE-based VR system. The desktop-
based system creates a semi-immersive VR environment for
stereoscopic visualisation, interaction, and quality analysis of
the product design. It is cost-effective and easy to operate, but
its users may be distracted by environmental disturbances that
could possibly diminish the efficiency of the product design
evaluation and improvement process. To alleviate disturbance
problems, the CAVE-based VR system provides an enclosed
room-like environment that blocks out disturbances, making it
possible for a design team to fully concentrate and collaborate
on evaluation and improvements of a product design. Hence,
the VP system enhances collaboration and communication of a
design team working on product development. It provides
effective tools to simulate and optimise MMLM processes that
fabricate prototypes for design evaluation and improvement to
facilitate subsequent product production. Through simula-
tions, validation of a product design can be readily iterated as
required without worrying about the manufacturing and
material cost of prototypes. Thus, key factors, such as product
shape and manufacturability, can be optimised accordingly.
Moreover, the resulting product design can be sent via the
Internet to customers for comments or marketing purposes.
The VP system therefore facilitates advanced product design
and helps reduce the time and cost of product development
considerably.
2. Review of related works
2.1. Multi-material layered manufacturing
Layered manufacturing (LM), also called rapid prototyping
(RP), has been widely used to produce prototypes of complex
shapes without tooling, particularly for manufacturing and
medical applications [11,12]. Multi-material layered manu-
facturing is an extension of the existing single-material LM
technology [13,14] for fabricating multi-material parts, such as
electronic products, advanced communication components,
drug delivery devices, and innovative cellular and cell-
containing tissue scaffolds [15–17]. A multi-material prototype
may be made of materials that change gradually from one type
to another, or of a collection of discrete materials. In contrast to
single-material ones, multi-material prototypes can differenti-
ate clearly one part from another, or tissues from blood vessels
of a human organ; and they perform better in rigorous
environments [18].
2.2. Virtual reality and virtual prototyping
VR systems have been successfully adopted for various
applications, especially military training, entertainment, surgi-
cal planning, manufacture simulation, marketing, and museum
exhibitions. Based on the level of immersion, VR systems are
either semi-immersive or full-immersive. A semi-immersive
VR system is typified by a desktop monitor or a pair of LCD
projectors with a large screen [4,7,9,19–21] on which
stereoscopic images are displayed, while a full-immersive
VR system is often characterised by a room-like CAVE
environment consisting of multiple screens. CAVE was first
developed by the Electronic Visualisation Laboratory (EVL) of
the University of Illinois-Chicago [22] to create a multi-person,
room-sized, 3D video and audio environment. Stereoscopic
images are projected onto three walls and the floor and are
viewed with active glasses equipped with a location sensor. The
following sections review some related works on the main
features and limitations of semi- and full-immersive VR
systems.
S.H. Choi, H.H. Cheung / Computers in Industry 59 (2008) 477–488 479
Morar and Macredie [19] pointed out that desktop-based VR
systems are relatively cheap and portable, and are popular for
3D interactive computer games, commercial and industrial
training, and modelling applications. Wang and Li [21]
proposed a desktop VR system for industrial training
applications, such as maintenance training for a refinery pump
system. They considered the system affordable, portable, and
easy-to-use, though it should be enhanced to provide better
realistic and interactive performance by upgrading the software
and hardware required.
Hoffmann et al. [9] reported that full-immersive VR
systems, such as the Powerwalls and the traditional CAVEs,
are often operated as VR centres or showrooms and are only
affordable by large enterprises due to the high cost and the
complexity and size of such installations. To overcome such
limitations, they attempted to develop a low-cost, compact, and
full-immersive VR system by extending the classical desktop
workplaces.
Fairen et al. [10] also pointed out that CAVE-based systems
do not fit in conventional offices due to special space
requirements. They proposed a low-cost, portable, and semi-
immersive VR system for cooperative inspection of complex
computer-aided designs. This portable system was intended for
demonstrations and presentations of designs at a client’s office.
Indeed, semi-immersive VR systems can be treated as an
affordable, easy-to-use, and convenient VR tool for stereo-
scopic visualisation, inspection, and interactions at locations,
such as in a customer’s office and at a trade fair, without the
need for large space and complex installation.
On the other hand, CAVE-based VR systems are a powerful
visualisation tool for collaborative applications, particularly in
military, medical, and automobile and aerospace industries. For
example, the USA army used a CAVE system to design, test,
and review new vehicle models before physical fabrication was
committed [23,24]. The development time and cost could be
significantly reduced because design errors were reduced as
communication between design team members was improved
through immersive visualisation of the digital vehicle prototype
in lieu of the physical ones. Indeed, CAVE provides a high
sense of immersion in real-time for multiple users, and the level
of immersion has a significant impact on user performance on
collaborative tasks [25].
However, high cost and complex installation hinder the
potential applications of CAVE-based VR systems in diverse
markets. To overcome such weaknesses, Li et al. [26]
developed a PC-based distributed multiple display VR system.
As programming of this system was based on the traditional C/
C++ language, it might not be easy to develop relatively
complex VR applications. Seron et al. [27] developed a CAVE-
like environment as a tool for full-sized train design. Although
Seron’s system could reduce design errors and streamline the
development of train products, the image quality, interactions,
and the level of immersion might need further improvement by
adding a floor with projected images at a cheaper cost.
From the discussions above, it is obvious that the desktop-
and the CAVE-based VR systems have different worthiness and
limitations. It would therefore be beneficial to integrate and
exploit the good features of the two systems. The following
section describes a versatile VP system for evaluation of
product designs and digital fabrication of multi-material
prototypes either in a semi-immersive environment or in a
full-immersive, disturbance-free environment to facilitate rapid
product development. This VP system provides flexibility for
users to choose either a desktop- or a CAVE-based VR platform
according to practical needs and available resources.
3. The proposed versatile virtual prototyping system
The proposed versatile VP system consists mainly of a suite
of software packages to simulate MMLM processes, including
colour STL modelling, slicing, topological hierarchy sorting of
slice contours for subsequent process planning, multi-toolpath
planning and generation, and build-time estimation [33–36]. In
particular, these packages are integrated with a set of control
modules and VR graphics kernels that drive both desktop- and
CAVE-based VR platforms to create semi- and full-immersive
visualisation of the MMLM processes at the user’s choice.
With the proposed VP system, designers can fabricate digital
multi-material prototypes, in lieu of costly physical ones, to
evaluate product designs and visualise the influences of critical
process parameters, such as build-direction, layer thickness,
and hatch space, on the MMLM process. The resulting digital
prototypes can be sent via the Internet to customers to solicit
comments, while the process parameters can be used for
optimal fabrication of physical prototypes. This approach
considerably reduces the number of costly physical prototypes
needed for rapid product development. Therefore, the
associated manufacturing overheads and product development
time can be reduced substantially, because digital prototypes
are mostly used and there is no worry about the cost and the
quality of physical prototypes.
Using the resulting set of optimal process parameters,
physical prototypes of desirable quality can be made quickly
and economically for detailed design evaluation. The physical
prototypes can also be used as master patterns for making tools
needed by conventional manufacturing processes, such as
injection moulding and CNC machining, for mass production
of the final products.
Furthermore, the VP system would be particularly useful for
small-batch production of customised products, which cannot
be produced with conventional processes economically.
Recently, LM has been widely explored for direct manufacture
of customised products. It is envisaged that when LM becomes
viable for direct manufacture of customised products, it will be
vital to validate the accuracy and quality of prototypes before
committing to physical fabrication. Therefore, the VP system
would be a practical simulation tool for rapid product
development.
Fig. 1 shows the flow of the VP system. Firstly, a product
model created by CAD or an MRI/CT digitiser is converted into
STL format, which is the industry de-facto standard. As STL is
monochrome or single-material, an in-house package is used to
paint the STL model, with each colour representing a specific
material.
Fig. 1. The flow of the versatile VP system.
S.H. Choi, H.H. Cheung / Computers in Industry 59 (2008) 477–488480
Secondly, a few steps are undertaken to prepare for
subsequent simulation of the MMLM process and visualisation
of the resulting digital prototypes: (a) slice the colour STL
model into a number of layers of a predefined thickness. If the
LM machine supports variable layer thickness, the STL model
may be sliced with an adaptive slicing algorithm to increase
fabrication efficiency. The resultant layer contours and material
information are stored in a modified Common Layer Interface
(CLI) file; (b) sort the slice contours with a contour sorting
algorithm to establish explicit topological hierarchy; (c) based
on the hierarchy information, multi-toolpath planning algo-
rithms are used to plan and generate multi-toolpaths by
hatching the slice contours with a predefined hatch space. The
hatch vectors are stored in the modified CLI file for fabrication
of digital prototypes and build-time estimation.
Thirdly, a versatile VR simulation system is used for digital
fabrication of multi-material prototypes. It allows users to
choose either a desktop- or a CAVE-based VR platform to
create a semi- or a full-immersive virtual environment,
respectively, for stereoscopic visualisation and quality analysis
of the resulting digital prototypes, with which product designs
can be reviewed and improved efficiently.
A suite of algorithms for LM process planning, such as
slicing, choice of build-direction, model orientation and layer
thickness, generation of sequential and concurrent multi-
toolpaths, and build-time estimation, are incorporated in the
proposed VP system. The details of these algorithms have been
presented in [28–36]. This paper focuses on the development of
the VR system. In particular, it addresses the enhancement of
versatility and effectiveness of virtual simulation for product
design and digital fabrication of multi-material prototypes at
affordable cost. The following section describes the desktop-
and the CAVE-based VR platforms in detail.
3.1. The desktop-based VR platform
The desktop-based VR system consists mainly of a software
package for stereoscopic visualisation of product designs and
Fig. 2. LCD projectors with a screen for semi-immersive VR display.
S.H. Choi, H.H. Cheung / Computers in Industry 59 (2008) 477–488 481
optimisation of MMLM processes. The software interfaces
with commercial desktop-based VR hardware to display a
model for stereoscopic visualisation. Using a desktop monitor,
which is relatively small but highly portable, a user wears a pair
of active shutter glasses that generate stereoscopic feelings by
synchronising with the display device to switch on and off the
images to the left eye and the right eye alternatively. This
creates a semi-immersive VR environment in which a designer
can stereoscopically visualise product designs and perform
quality analysis. If a much larger display is needed, a pair of
LCD projectors with a large screen as in Fig. 2 can be used. A
user wears a pair of oppositely polarised glasses that filter the
polarised images for the left eye and the right eye, respectively.
With this wall-sized screen, a group of designers can participate
in stereoscopic visualisation and collaborative review of
product designs in the semi-immersive VR environment. This
indeed improves exchange of ideas among a design team.
In addition, the software package consists of a Product
Viewer module and a Virtual Prototype Fabrication module,
based on the WorldToolkit (WTK) graphic libraries, for
simulation of the MMLM process. The Product Viewer module
displays a colour product model in a semi-immersive VR
environment in which a small group of designers can work
together to study and improve the product design; the Virtual
Prototype Fabrication module can then fabricate digital multi-
material prototypes of the product.
A dexel-based approach is adopted for digital fabrication of
prototypes [28]. A dexel is a hatch vector representing the path
that a tool has to follow within a contour to build a portion of a
layer. By building a volume of a specific height and a width
around a dexel, a strip of material may be represented. Hence,
rectangular solid strips are laid to form a layer, which is
subsequently stacked up to form a prototype during a digital
fabrication process.
During the fabrication process, a designer can observe how
a prototype is fabricated. Once it is finished, the resulting
digital multi-material prototype can be studied using the
utilities provided to visualise the quality of the prototype that
the LM machine will subsequently deliver. The designer can
navigate around the internal and opaque structures of the
prototype to investigate the design. Besides, the colour STL
model can be superimposed on its digital prototype for
comparison, with the maximum and the average cusp
highlighted to indicate the dimensional deviations. A tolerance
may be set to highlight locations with deviations beyond the
limit. The designer may thus identify and focus on the parts that
would need modifications. To improve the accuracy and the
surface quality of some specific features of the prototype, the
process parameters, such as the build-direction, the model
orientation, the layer thickness, and the hatch space, may be
tuned accordingly.
After the visualisation process, the colour STL model of the
toy car is sliced, for example, into 120 layers with a thickness of
0.194 mm, because most current LM machines support only
uniform layer thickness during a prototype fabrication process.
If the LM machine to be simulated supports variable layer
thickness, the model may be instead sliced with an adaptive
slicing algorithm. The resulting layer contours are then sorted
to establish the topological hierarchy for generation of multi-
toolpaths with a hatch space of 0.400 mm. Subsequently, the
Virtual Prototype Fabrication module fabricates a digital
prototype by depositing the rectangular solid strips one by one
at an appropriate z-height, as shown in Fig. 3. The resulting
virtual prototype of the toy car can be manipulated for visual
inspection, as in Fig. 4. Furthermore, it can be superimposed on
its STL model to highlight the surface texture and the
dimensional deviations, as in Fig. 5. The system also calculates
the cusp heights to evaluate the overall dimensional deviations.
In this example, the average and the maximum cusp heights are
0.098 and 0.164 mm, respectively. Suppose that any deviations
more than 0.170 mm are considered unacceptable, the designer
may choose to highlight the areas which are out of the design
limit for subsequent investigation of these critical features.
Excessive deviations are highlighted with red or green pins. The
red pins point to the maximum deviations whereas the green
ones point to unacceptable deviations. If unsatisfactory
deviations are located at important parts of the model, the
designer may choose either to change the model orientation to
shift the deviations or to reduce the layer thickness and the
hatch space to improve the accuracy. When it is necessary to
Fig. 3. Digital fabrication process of a toy car prototype.
S.H. Choi, H.H. Cheung / Computers in Industry 59 (2008) 477–488482
assess detailed assembly fitness of the various parts of the toy
car, the parts can be stored as individual STL models for quality
analysis and digital fabrication. Digital fabrication of a part
prototype can be repeated until a set of acceptable process
parameters are obtained. Subsequently, physical prototypes of
all parts are produced and assembled to form a complete toy car
prototype. Furthermore, the physical prototypes can be
processed and used as master patterns to make tools for mass
production of the product.
Therefore, the proposed desktop-based VR system is an
easy-to-use and cost-effective tool for visualisation and digital
fabrication of multi-material prototypes to facilitate product
design review and improvement. However, the semi-immersive
VR environment may be susceptible to environmental
disturbances, diminishing the designers’ true feeling and
concentration and hence their efficiency in the design process.
Fig. 4. Two perspectives of
To address this problem, the level of immersion is enhanced
by integrating the VP system with a CAVE-based VR system
with multiple screens to provide a full-immersive virtual
environment for vivid stereoscopic visualisation and interaction
in a natural way. As such, a design team can fully immerse in
exploration, study, and improvement of a product design,
including assemblies, sub-assemblies, and components, well
before they ever exist physically in reality. Hence, the time and
cost of product development can be further reduced.
3.2. The CAVE-based VR platform
The CAVE-based VR platform consists of a cluster of PCs
with a cubicle of three walls on a floor. An immersive virtual
environment is created by projecting stereoscopic images on
three 10ft � 8ft screens on the walls, namely the front, the right,
the toy car prototype.
Fig. 5. Superimposition of the toy car prototype on its STL model to highlight excessive materials and dimensional deviations.
S.H. Choi, H.H. Cheung / Computers in Industry 59 (2008) 477–488 483
and the left, respectively, and on a 10ft � 10ft screen on the
floor. Fig. 6a and b shows, respectively, the architecture and the
physical construction of the PC-based CAVE system, called
imseCAVE, in the IMSE Department at the University of Hong
Kong.
Each projection screen has a reflector and two LCD
projectors controlled by two related PCs. The LCD projectors
are specially designed with polarising lenses to produce high-
resolution stereoscopic images. A VR engine, consisting of a
cluster of network PCs, coordinates the projectors to project
Fig. 6. (a) The architecture of the imseCAVE. (
images on the related screens to create an immersive virtual
environment. This configuration forms a relatively low-cost,
configurable, and flexible CAVE-based VR system, which can
be conveniently integrated to form the proposed versatile VP
system to facilitate product development. The hardware is
controlled by a software package, which can be separated into
three layers, as shown in Fig. 7.
The bottom layer includes basic system software and
hardware, such as Windows XP OS and graphics kernel for
control of the PC network, interface devices, and projectors.
b) Physical construction of the imseCAVE.
Fig. 7. Software architecture of the PC cluster-based CAVE system.
S.H. Choi, H.H. Cheung / Computers in Industry 59 (2008) 477–488484
The middle layer is a PC cluster-based library that
coordinates all operations of the CAVE system. It synchronises
all the devices to create correct perspective for each screen,
keeps track of which screens are in use, and provides the
applications with the current states of all the CAVE elements. It
consists of three sub-systems: (i) I/O device sub-system for
controlling I/O devices; (ii) display subsystem for projecting
images on the corresponding screens; and (iii) the network
subsystem for keeping communication and synchronisation
between all clustering PCs.
The top layer is a package of tailor-made application
programs, developed with the PC cluster-based CAVE library,
for immersive visualisation and simulation in the CAVE
system. This application package is implemented in an object-
oriented programming tool, called the Virtools Dev, and its add-
on library, called Virtools VR Pack [37]. The Virtools Dev
contains a suite of algorithms to help programmers create,
visualise, manipulate, and track objects in the virtual
environment. Besides, the Virtools VR Pack allows users to
tailor applications for producing full-immersive, life experi-
ences using the PC-based distributed computing.
In using the Virtools Dev tool, a Script is the visual
representation of a behaviour applied to an element. A
behaviour is described with Behaviour Building Blocks (BBs)
which are a visual representation of a software element
known as a function. The software provides a collection of
pre-defined BBs that enable users to create an application
script conveniently. Scripts are performed by the Behaviour
Engine.
With this approach, the application package can be
conveniently developed for stereoscopic visualisation and
simulation of the MMLM process in an immersive CAVE
virtual environment. The application package for the PC
cluster-based CAVE system contains two modules, namely,
Product Viewer and Virtual Fabricator, similar to those for the
desk-based system.
The Product Viewer displays a virtual product in a CAVE
virtual environment in which users can fully immerse to
manipulate and study the design, as in Fig. 8. It also facilitates
manipulations of a virtual product, including rotation, and scale
up/down, toggling visibility/invisibility of a component, using
wireless I/O devices, such as a mouse, a keyboard, and a
joystick. The designer can hide the external car body to study
the gearbox assembly from different perspectives.
Such manipulation functions are developed using the
Virtools Dev. A product model created by the CAD software
or the digitised equipment is first converted into the VRML file
format, and then imported to the Virtools Dev to add functions
and behaviour of each product component by linking the BBs
accordingly. To develop functions for rotating and scaling up/
down the virtual product via a number of specific keys of a
wireless keyboard, three standard BBs, called Switch On Key,
Rotate, and Scale, are used. When a user presses down a
particular key for rotating the model, the Switch On Key BB is
triggered and a output signal is sent to the corresponding Rotate
BB to activate the rotation behaviour. Thus, the context of the
Product Viewer module can be easily developed.
In addition, the Virtual Fabricator, similar to that in the
desktop-based system, is created with the Virtools Dev for
digital fabrication of multi-material prototypes. When the
fabrication completes, designers can fully immerse in the
CAVE virtual environment for stereoscopic visualisation and
quality analysis of the resulting multi-material virtual proto-
type. This full-immersive environment blocks out most
disturbances and hence enhances the efficiency of the design
review and improvement process.
For the Virtual Fabricator to simulate a prototype fabrication
process, the multi-toolpaths are translated into a dataset
supported by the Virtools Dev for the Virtual Fabricator to load
and fabricate. The Virtual Fabricator has two scripts created
with the Virtools Dev, one for loading the dataset, namely
DataLoad, and another for building 3D models for simulation
of the prototype fabrication process, namely Build3DModel.
The script DataLoad consists of two BBs, called Array Load
for loading the tabular data in cells into an array from the
formatted file and Activate Script for activating a script,
respectively. These two BBs are linked together with behaviour
links (bLinks). Each BB has its own parameters. When the
DataLoad script finishes the loading of data, the Activate Script
BB is triggered to activate the Build3DModel script via a
behaviour link (bLink). Then, the Build3DModel starts
simulating the prototype fabrication process layer by layer.
The Virtual Fabricator also adopts the dexel-based approach
for simulation of the prototype fabrication process, as in the
desktop-based system. It uses an object Cube, which is a
standard 3D entity data resource and is stored in NMO file
format, to represent a dexel with its specific position, size, and
material property. Hence, three BBS, namely Set Position,
Scale, and Set material, are used to represent the position, size,
and material of a dexel based on the data stored in a text file. In
addition, the Iterator BB is used to control the layer-by-layer
loop process.
For the scripts of both the Product Viewer and the Virtual
Fabricator to function properly in the CAVE-based system,
they have to be integrated with a script program called the PC
cluster-based CAVE Coordinator. The PC cluster-based CAVE
Coordinator is created with the Virtools VR pack, which
provides a set of VR libraries containing a package of standard
BBs for users to develop applications needed to control a
cluster of PCs and projectors and to generate a full-immersive
virtual environment. The Coordinator can be treated as a
Fig. 8. Study of a product design in the CAVE virtual environment.
S.H. Choi, H.H. Cheung / Computers in Industry 59 (2008) 477–488 485
middle layer between the Product Viewer/the Virtual
Fabricator modules and the PC cluster-based CAVE display
to coordinate the operations of the whole CAVE-based VR
system. It synchronises a cluster of PCs to distribute image
signals to the related projectors for projection on the screens,
and hence the creation of an immersive virtual environment
through the network system. In this cluster of PCs, one is
master for receiving user input signals while the remaining
ones are slaves for screen display.
Firstly, the master receives the peripheral state of the
joystick, the keyboard, and the mouse, etc. and then sends this
state to each slave. Secondly, the slaves wait for this state to
arrive and acknowledge reception to the master. Based on the
same shared causes, they will all compute their own frame
when every PC holds the shared state. Thus, using the Virtools
VR Pack, this distributed computing technique can be easily
Fig. 9. Digital fabrication of
developed by logically linking the Virtools VR BBs, such as VR
Host Id and VR Distrib, to develop specific applications.
It can be seen that the PC cluster-based CAVE system above
is relatively convenient and flexible, making full-immersive VR
a versatile and affordable tool for small-and-medium sized
companies to develop products.
4. A case study
Hong Kong produces a wide range of footwear products
mainly for export. In recent years, many footwear companies
have attempted to develop their own brands by improving
capabilities of product design, tooling, and quality control,
while some others focus on providing tailor-made services to
take care of the special needs of customers. The proposed
versatile VP system would therefore be useful for the footwear
the shoe sole prototype.
Fig. 10. Digital fabrication process of a multi-material shoe sole prototype.
Fig. 11. Superimposition of the shoe sole prototype on its STL model.
S.H. Choi, H.H. Cheung / Computers in Industry 59 (2008) 477–488486
industry to help save product development cost and time. In the
following section, a casual shoe sole, which affects the comfort
and performance of the footwear, is used to demonstrate a
possible application of the proposed versatile VP system.
4.1. A casual shoe sole
Using the VP system, a shoe designer can choose a desktop-
based VR system to review the shape, colour, and ergonomics
of the shoe sole in a semi-immersive virtual environment. But to
minimise disturbances and to increase the level of immersion to
help a design team focus on exchange of ideas for design
improvements, a CAVE VR system can be used instead. This
helps reduce product development cost and time substantially
since potential errors can be avoided in the early design stage.
After visualisation and evaluation of the design, the shoe
sole model in STL format is firstly sliced. In this case, it is
sliced into 88 layers with a layer thickness of 0.178 mm.
Secondly, multi-toolpaths are planned and generated by
hatching each layer contours with a hatch space of
0.496 mm. Subsequently, based on the resulting multi-toolpaths
containing geometric and material information, a virtual multi-
material shoe sole prototype is fabricated, either in a semi- or
Fig. 12. (a) Measurement of foot plantar pressure using an F-Scan1 system. (b) Foot plantar pressure profiles for the pair of physical shoe soles.
S.H. Choi, H.H. Cheung / Computers in Industry 59 (2008) 477–488 487
full-immersive environment, as shown in Fig. 9. During the
digital fabrication process, the designers can visualise how it is
fabricated layer by layer to reveal the detail structure, as shown
in Fig. 10.
The resulting shoe sole prototype can be superimposed on its
colour STL model to highlight dimensional deviations, and the
overall dimensional deviations can be analysed accordingly, as
in Fig. 11. In this case, the average and the maximum cusp
heights are 0.092 and 0.179 mm, respectively, and the regions
with deviations exceeding 0.160 mm are highlighted. The red
pins point to the maximum deviations whereas the green ones
point to unacceptable deviations.
With this result clearly visualised, the designer may choose
to modify the design or change a new set of process parameters
to minimise the dimensional deviations. When the quality is
deemed acceptable, the process parameters can be used for
subsequent fabrication of physical prototypes. Hence, the
MMLM process is optimised and the number of costly physical
prototypes reduced accordingly. As such, a pair of physical
shoe sole prototypes of elastomeric material with rubber-like
properties [38] can be fabricated on a 3D printing machine to
test ergonomic fitness by measuring the profiles of the plantar
pressure induced by a user’s feet. To do this, an F-Scan1 system
with a pair of paper-thick sensors [39], as shown in Fig. 12a,
may be put on the shoe soles for the user to step on for testing.
By studying the plantar pressure profiles generated as shown in
Fig. 12b, the designer can evaluate the ergonomic fitness of the
shoe soles quantitatively, and modify the design accordingly, if
deemed necessary.
The design modification-evaluation-testing process above
can be repeated quickly until a pair of shoe soles of satisfactory
design is obtained. Therefore, the proposed versatile VP system
is useful for reducing the cost and time of product development.
5. Conclusion
This paper proposes a versatile VP system which integrates
the good features of semi- and full-immersive VR to enhance
the versatility and effectiveness of virtual simulation for
product design and digital fabrication of multi-material
prototypes at affordable cost.
The VP system comprises mainly a suite of software
packages for simulation of MMLM processes, including multi-
toolpath planning, build-time estimation, and accuracy
analysis. It can drive a desktop-based VR system with either
a monitor or a large non-depolarising screen to generate a semi-
immersive VR environment, which is cost-effective, portable,
and easy to operate, for review and improvement of product
designs. To minimise environment disturbances and to enhance
the level of immersion, the VP system can control a PC cluster-
based CAVE system to create a full-immersive VR environment
that enhances collaboration and communication of a design
team working on product development. It is indeed an effective
and versatile tool for rapid product development to meet ever-
increasing diversities of customer demands, stringent cost
control, and complexity of new products.
Acknowledgements
The authors would like to acknowledge the Research Grant
Council of the Hong Kong SAR Government and the CRCG of
the University of Hong Kong for their financial support for this
project.
References
[1] H.T. Kim, J.K. Lee, J.H. Park, B.J. Park, D.S. Jang, Applying digital
manufacturing technology to ship production and the maritime environ-
ment, Integrated Manufacturing Systems 13 (5) (2002) 295–305.
[2] G. Wohlke, E. Schiller, Digital planning validation in automotive industry,
Computers in Industry 56 (4) (2005) 393–405.
[3] M. Kerttula, T. Tokkonen, Virtual design of multiengineering electronics
systems, IEEE Computer 34 (11) (2001) 71–79.
[4] T.S. Mujber, T. Szecsi, M.S.J. Hashmi, Virtual reality applications in
manufacturing process simulation, Journal of Materials Processing Tech-
nology 155/156 (2004) 1834–1838.
[5] A. Gomes de Sa, G. Zachmann, Virtual reality as a tool for verification of
assembly and maintenance processes, Computers & Graphics 23 (3)
(1999) 389–403.
[6] G.M. Bochenek, J.M. Ragusa, Study results: the use of virtual environ-
ments for product design, in: Proceedings of IEEE International Con-
ference on Systems, Man, and Cybernetics, Hyatt Regency La Jolla, San
Diego, California, USA, 11–14 October, (1998), pp. 250–1253.
[7] M. Weyrich, P. Drews, An interactive environment for virtual manufactur-
ing: the virtual workbench, Computers in Industry 38 (1) (1999) 5–15.
S.H. Choi, H.H. Cheung / Computers in Industry 59 (2008) 477–488488
[8] H. Patel, S. Sharples, S. Letourneur, E. Johansson, H. Hoffmann, L.
Lorisson, D. Saluaar, O. Stefani, Practical evaluations of real user
company needs for visualization technologies, International Journal of
Human-Computer Studies 64 (3) (2006) 267–279.
[9] H. Hoffmann, O. Stefani, H. Patel, Extending the desktop workplace by a
portable virtual system, International Journal of Human-Computer Studies
64 (3) (2006) 170–181.
[10] M. Fairen, P. Brunet, T. Techmann, MiniVR: a portable virtual reality
system, Computers & Graphics 28 (2) (2004) 289–296.
[11] N. Hopkinson, R.J.M. Hague, P.M. Dickens, Rapid Manufacturing: An
Industrial Revolution for the Digital Age, John Wiley & Sons, 2006.
[12] I. Gibson, Advanced Manufacturing Technology for Medical Applica-
tions: Reverse Engineering, Software Conversion and Rapid Prototyping,
John Wiley & Sons, 2005.
[13] M.A. Jafari, W. Han, F. Mohammadi, A. Safari, S.C. Danforth, N.
Langrana, A novel system for fused deposition of advanced multiple
ceramics, Rapid Prototyping Journal 6 (3) (2000) 161–175.
[14] K. Lappo, B. Jackson, K. Wood, D.L. Bourell, J.J. Beaman, Discrete
multiple material selective laser sintering (M2SLS): experimental study of
part processing, in: D.L. Bourell, et al. (Eds.), Solid Freeform Fabrication
Symposium, The University of Texas, Austin, Texas, (2003), pp. 109–
119.
[15] A. Safari III, J. Cesasrano, P.G. Clem, B. Bender, Fabrication of advanced
functional electroceramic components by layered manufacturing (LM)
methods, in: Proceedings of the 13th IEEE International Symposium on
Applications of Ferrolelectronics, Nara, Japan, (2002), pp. 1–6.
[16] E. Sachlos, N. Reis, C. Ainsley, B. Derby, J.T. Czernuszka, Novel collagen
scaffolds with predefined internal morphology made by solid freeform
fabrication, Biomaterials 24 (8) (2003) 1487–1497.
[17] B. Starly, W. Lau, T. Bradbury, W. Sun, Internal architecture design and
freeform fabrication of tissue replacement structures, Computer-Aided
Design 38 (2) (2006) 115–124.
[18] K. Samanta, B. Koc, Feature-based and material blending for free-form
heterogeneous object modelling, Computer-Aided Design 37 (3) (2005)
287–305.
[19] S.S. Morar, R.D. Macredie, Usability issues of desktop virtual environ-
ment applications, Virtual Reality 7 (3/4) (2004) 175–176.
[20] W.S. Huang, C.Y. Shing, C.P. Fung, T.Y. Chuang, M.C. Jeng, J.L. Doong,
Assessment of various display devices on a virtual reality-based hand
rehabilitation system, Robotica 22 (1) (2004) 9–14.
[21] Q.H. Wang, J.R. Li, A desktop VR prototype for industrial training
applications, Virtual Reality 7 (3/4) (2004) 187–197.
[22] CAVE, http://evlweb.eecs.uic.edu/core.php?mod=4&type=1&indi=161
Electronic Visualisation Laboratory, 2005.
[23] G.M. Bochnek, J.M. Ragusa, Virtual (3D) collaborative environments: an
improved environment for integrated product team interaction? in: R.H.
Sprague (Ed.), The 36th International Conference on System Sciences,
IEEE Computer Society, Big Island, Hawaii, 2003.
[24] H. Creagh, CAVE automatic virtual environment, in: IEEE Proceedings of
Electrical Insulation Conference and Electrical Manufacturing & Coil
Winding Technology Conferences, Indianapolis, Indiana, 23–25 Septem-
ber, (2003), pp. 499–503.
[25] M. Narayan, L. Waugh, X.Y. Zhang, P. Bafna, D. Bowman, Quantifying
the benefits of Immersion for Collaboration in virtual environments, in:
Proceedings of the ACM Symposium on Virtual Reality Software and
Technology, Monterey, California, USA, 7–9 November, 2005.
[26] W.J. Li, C.C. Chang, K.Y. Hsu, M.D. Kuo, D.L. Way, A PC-based
distributed multiple display virtual reality system, Displays 22 (5)
(2001) 177–181.
[27] F.J. Seron, D. Gutierrez, J.A. Magallon, E.J. Sobreviela, J.A. Gutierrez, A
CAVE-like environment as a tool for full-size train design, Virtual Reality
7 (2) (2004) 82–93.
[28] S.H. Choi, S. Samavedam, Visualisation of rapid prototyping, Rapid
Prototyping Journal 7 (2) (2001) 99–114.
[29] S.H. Choi, S. Samavedam, Modelling and optimisation of rapid proto-
typing, Computers in Industry 47 (1) (2002) 39–53.
[30] S.H. Choi, K.T. Kwok, A tolerant slicing algorithm for layered manu-
facturing, Rapid Prototyping Journal 8 (3) (2002) 161–179.
[31] S.H. Choi, A.M.M. Chan, A dexel-based virtual prototyping system
for product development, Rapid Prototyping Journal 8 (5) (2002) 300–
314.
[32] S.H. Choi, K.T. Kwok, A topological hierarchy-sorting algorithm for
layered manufacturing, Rapid Prototyping Journal 10 (2) (2004) 98–113.
[33] S.H. Choi, H.H. Cheung, A multi-material virtual prototyping system,
Computer-Aided Design 37 (1) (2005) 123–136.
[34] S.H. Choi, H.H. Cheung, A multi-material virtual prototyping system for
product development and biomedical engineering, Computer-Aided
Design and Applications 2 (1–4) (2005) 329–338.
[35] S.H. Choi, H.H. Cheung, A topological hierarchy-based approach to
toolpath planning for multi-material layered manufacturing, Computer-
Aided Design 38 (2) (2006) 143–156.
[36] S.H. Choi, H.H. Cheung, Multi-material virtual prototyping for product
development and biomedical, Computers In Industry 58 (5) (2007) 438–
452.
[37] Virtools, http://www.virtools.com, 2006.
[38] Materials, http://www.zcorp.com/products/materials.asp, 2007.
[39] F-Scan1, http://www.tekscan.com, 2007.
S.H. Choi is associate professor in the IMSE Depart-
ment at the University of Hong Kong. He obtained
both his BSc and PhD degrees at the University of
Birmingham. He worked in computer industry as
CADCAM consultant before joining the University
of Hong Kong. His current research interests include
CADCAM, advanced manufacturing systems and
virtual prototyping technology.
H.H. Cheung gained his BEng degree from the
IMSE Department at the University of Hong Kong.
He continued his postgraduate research study in the
Department, and his research interest is in virtual
prototyping technology.