Middleware Communication System for Fixture based Flexible
Manufacturing Systems
N. Chiraga, A. Walker and G. Bright
Discipline of Mechanical Engineering, University of KwaZulu-Natal, Durban, South Africa
Keywords: Distributed Flow Control, Wireless Communication in Manufacturing, Middleware, Mass Customization.
Abstract: Today’s factory floors are a distributed system of heterogeneous components. They exist a high degree of
production customization to cater for this most factory floor systems are being adapted for customization.
Most current manufacturing communication systems do not have a flexible architecture to serve the
changeable and dynamic market needs thus presenting a research gap in flexible communication. This paper
discusses a middleware communication system that allows for flexible optimal control of information flow in
a fixture based production system. The architecture of the system aims to integrate seamlessly a heterogeneous
factory environment, for autonomous monitoring and control of information on the factory floor.
1 INTRODUCTION
The manufacturing industry is currently experiencing
a paradigm shift into the Fourth Industrial Revolution
in which customers are increasingly at the epicentre
of production (EFFRA, 2013). This revolution has
prompted major research into factories of the future;
forcing industries to re-examine their systems (Bloem
et al., 2014). Schwab et al stated in the recent world
economic forum stated that the Fourth Revolution is
evolving at an exponential rather than linear pace
(Schwab, 2016).The high degree of production
customization and personalization requires a flexible
manufacturing system that will rapidly respond to the
dynamic and volatile changes driven by the market
(Qiao et al., 2000). Mass customization has seen
limited adoption, thus giving raise to research into the
implementation of manufacturing systems that are
highly responsive to these rapid changes in market
demands. They is a gap in technology that allows for
optimal flow of information and optimal
manufacturing operations on the shop floor regardless
of the rapid changes in fixture and part demands. The
mass customisation manufacturing (MCM) paradigm
has created a problem in manufacturing control
implementations, as each customer has the potential
to disrupt production operations resulting in
downtime (Walker and Bright, 2013). Factory
communication systems now also have to cater for the
high degree of production customization and
personalization.
It is essential to attain at least 99.99% uptime or
better on the shop floor during these rapid changes
(Qiao et al., 2000). Depending on the size and nature
of business, indirect costs due to downtime can range
from tens of thousands to hundreds of thousands of
dollars (Qiao et al., 2000). A reliable advanced
factory communication system architecture is critical
for optimal network performance. A deep and
efficient integration of information flow, information
analysis, and customer input in the production
network is necessary. There is a need to provide the
right information at the right time and show the
manufacturing decision maker how current
conditions on the plant floor can be optimised to
improve customized production output. Salvador et
al. identified three fundamental capabilities
determining the ability of a factory to mass-customize
its offerings, i.e. solution space development, robust
process design, and choice navigation (Salvador et
al., 2009).
This paper aims to develop the robust process
design capability in Factory Communication
Systems; the capability to create, reuse, or recombine
existing factory shop fixture resources to fulfil a
stream of differentiated customer needs (Salvador et
al., 2009). Flexible automation, process modularity,
and adaptation are approaches that can be taken to
develop robust process design. Flexible automation
can be described as automation that is not fixed or
rigid and can handle the customization products.
While process modularity is the segmentation of
146
Chiraga, N., Walker, A. and Bright, G.
Middleware Communication System for Fixture based Flexible Manufacturing Systems.
DOI: 10.5220/0005972201460152
In Proceedings of the 13th International Conference on Informatics in Control, Automation and Robotics (ICINCO 2016) - Volume 1, pages 146-152
ISBN: 978-989-758-198-4
Copyright
c
2016 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
existing organizational and factory shop resources
into modules that can be reused or recombined to
fulfil differentiated customers’ needs (Salvador et al.,
2009).
To facilitate the flexible communications between
the high-level software systems Enterprise Resource
Planning (ERP), SCADA and Manufacturing
Execution System (MES)) and users (control and
industrial engineers), and the heterogeneous factory
floor environment middleware technology is needed.
This paper details an advanced Factory Middleware
Communication System (FMCS) that uses a
middleware communication system to allow for
flexible control and information exchange in a factory
environment driven by the dynamic customer needs
for production execution. The aim of this study was
to seamlessly integrate manufacturing floor
communication with the customers’ decisions in
product configuration.
2 AN OVERVIEW OF THE
FACTORY MIDDLEWARE
COMMUNICATION SYSTEM
(FMCS)
Traditional Supervisory Control and Data
Acquisition (SCADA) systems collect data from
various sensors nodes deployed in remote locations
and then transmit it to a central controller which then
manages and controls this data ("SCADA").While the
Manufacturing Execution System(MES) tracks and
documents the transformation of raw materials
through finished goods("Manufacturing Execution
System"). The Enterprise Resource Planning (ERP)
tracks the factory’s resources, i.e. raw materials
supply and production capacity ("Enterprise
Resource Planning"). This Factory Middleware
Communication system is an advanced
manufacturing strategy aiding the SCADA system,
MES and ERP; and equipping them with intelligent
handling capabilities for the rapidly changing
customer needs, which in turn result in the production
processes experiencing rapidly varying fixture and
part demands. The figure below illustrates where the
developed system exits in a manufacturing pyramid.
Figure 2 shows the general overview of the
system. The system solution copes with a demand
characterized by at least a subset of the following
properties: low-mid demand volumes, mid-high
variety of the part mix, short product lifecycle, and
mid-high customization. The following sections
explain each module in the system.
Figure 1: Manufacturing production levels.
Figure 2: Factory Middleware Communication System
(FMCS) Overview.
2.1 State Detection Module
The FMCS operates using state detection modules
(SDM) attached to each factory cell to determine the
state of each cell at any given time on the production
floor. The distributed SDMs; microcontroller-based
modules form a sensor network, with state detectation
capabilities. In this research, we are primarily
interested in the information presented at each factory
cell. The aim is to create an information intense
environment rather than just transmit instructions.
Information is the reduction of uncertainty; it gives
meaning and context on the state of each cell. In
information theory Shannon describes entropy as the
average value of information contained by a system
(Lesne, 2011).
Suppose that a discrete variable X, has n possible
outcomes;
,
,
,…..,
the probability for each
outcome being
. Shannon entropy is defined as:

log

(1)
Middleware Communication System for Fixture based Flexible Manufacturing Systems
147
Where:
0 And
1

The variable denotes a system, which is the cell in
our case,
and
1,2,…., are its n possible
states and their probabilities in the cell respectively.
The amount of information needed to describe the
cell, is  i.e. information entropy (Lesne, 2011).
Considering a factory cell X on a factory floor, six
states can be defined;
i. Optimal working state
ii. Running slow state
iii. Idle state
iv. Scheduled maintenance state
v. Unscheduled maintenance state
vi. Downtime state
Based on the probabilities of each state, using
Shannon Theorem the entropy for each cell can be
calculated and hence the overall entropy of the entire
information system.
2.2 The State Communication Module
(SCM)
The SCM facilitates the communication link between
the state detection modules and the central control
computer, which runs the Central FMCS Control
(CFC) software.
Communication techniques range from wired to
wireless communication protocols. Publications have
shown emerging work ushering in the wireless
communication protocols into industrial networks.
(Buda et al., 2010). To keep up with technological
trends wireless technology was used for the
development of the communication network.
Wireless technology offers flexibility; modules can
be introduced to the system in a plug and play manner
without having to worry about re-wiring. Factory
propagation environments however, have many
metallic surfaces and moving objects, which can be
characterised as harsh conditions that can affect the
operational quality of radio based communication
systems (Buda et al., 2010). Wireless communication
is implemented with ZigBee wireless protocol, a
well-adapted technology for industrial applications.
ZigBee operates at 2.4GHz frequency, which lays in
the bandwidth of frequency not affected by the
interference in factories. (Buda et al., 2010). ZigBee
is based on an IEEE 802.15.4 standard. XBee Series
2 are used for the Zigbee wireless communication.
The SCMs help us achieve process modularity; state
information is continually transmitted wirelessly to
the central control.
2.3 Central FMCS Control (CFC)
2.3.1 Middleware Communication Layer
Today’s factory floors are a distributed system of
heterogeneous components. There exist a wide range
of applications and system platforms, making
heterogeneity in factories predominant. Each SDM
stands out as a heterogeneous component in a cellular
configuration. This is where middleware technology
comes in; middleware masks heterogeneity on the
factory floor and allows for flexible control and
exchange of information. Using a thick software-
intensive middleware layer negates the need for a
hardware supported middleware system in the control
system. The objective is maintain optimal
manufacturing operations on the shop floor during
model definition changes of customer-desired
products. Internet Communications Engine (Ice) is a
modern object-oriented middleware platform that
enables you to build distributed applications with
minimal effort (Roulet-Dubonnet et al., 2013).
With Ice, there is no need to worry about details
such as opening network connections, serializing and
deserializing data for network transmission, or
retrying failed connection attempts. Its features
include location independence; the client does not
need to know of the specifics of the target object’s
address, platform independence; multi-vendor
manufacturing devices can integrate seamlessly,
programming language independence; does not
matter if client and server processes are implemented
with the same language or different languages. The
middleware layer works as the workflow manager,
the nervous system of the network created allowing
for interoperability and seamless flow of information.
Figure 3 shows the middleware client; sever
structure (Roulet-Dubonnet et al., 2013).The CFC
acts as the client and the SCMs as the server.
Figure 3: Ice Client and Server Structure.
When a client i.e. central control invokes an
operation, the following steps take place:
i. Locates target object i.e. the SCMs located at
each cell in the network
ICINCO 2016 - 13th International Conference on Informatics in Control, Automation and Robotics
148
ii. Activates the server application in the SCMs,
if the server is not already running
iii. Transmits any arguments for the call to the
object via the ZigBee wireless network
iv. Waits for request to complete
v. Returns any out parameters or return value to
the client when a call completes successfully.
The return parameters in our case is the state
information gathered by the SDMs, fixture
information from the fixture reconfiguration
system and machine orientation from the
reconfigurable machine system.
vi. Returns an exception to the client when a call
fails
2.4 Intelligent Processing and Program
Assignment in the Central Control
(CFC)
The CFC is the middleware and intelligence layer of
the system, the workflow manager. It continually
polls for information from the factory floor and stores
it into a database. Based on the factory floor state,
fixture configuration, availability, and the machine
configuration an order is processed and control
instructions are assigned to the work cells. Through
this intelligence of raw factory floor state data fed by
the SCMs, an information-based configuration is set-
up. Sequences of job arrive, entering the shop floor
input buffer and are logged in through the graphical
user interface and order parameters are noted
according to the customers’ decisions in product
configuration. Jobs are dealt with using the MIMO
queue concept. The product specifications make up
the procedural rules for the workflow decision tree.
The ability to effectively communicate the factory
floor state and the characteristic response of the
system to a new product configuration are of utmost
importance in this research. The EPR keeps track of
the available raw materials i.e. fixture database on the
factory, if a new product configuration requires a new
fixture, instructions are sent to the 3D Printing cell to
produce a new fixture. Once the fixture production is
done, the fixture database is updated and the
production plan and control instruction are sent to
each cell on the factory floor. Figure 4 below
explains the procedural steps in the intelligent process
up until the order program is assigned to the factory
floor for manufacture.
Figure 4: Procedural steps in the intelligent processing of
an order.
3 EXPERIMENTATION ON THE
SYSTEM
The system was implemented and tested in the UKZN
Manufacturing and Mechatronics Laboratory. The
laboratory layout depicts a factory floor environment;
it is made up of a cellular configuration. The cellular
make-up of the configuration includes; a Material
handling cell, 3D Printer cell, Assembly cell,
manufacturing cell and Quality control cell. State
detection modules as well as state communication
modules were placed in each cell, forming a star
network. The CFC software layers links up with two
pre-existing systems in the laboratory i.e.; the
machine reconfiguration system and the fixture
reconfiguration system. The machine reconfiguration
system provides the CFC with the physical
configuration of the machines on the factory floor and
allows the machines to be reconfigured accordingly.
While the fixture reconfiguration system provides the
current fixture configuration at any given time and
again allows for fixture reconfiguration.
The flexibility of this system is random-order,
there are substantial variations in part configurations,
new part designs are continually being introduced to
the system. It is a recognition that a factory cannot
do everything; it must limit its options to a certain
Middleware Communication System for Fixture based Flexible Manufacturing Systems
149
scope of products activities in which it can best
compete.
Table 1: Hypothetical initial scope of product variety.
Hard Product Variety Soft Product Variety
Product 1 5 variations
Product2 3 variations
3.1 Aim
The experiment associated with the situation
described above aimed to prove that it was possible,
using the technology developed in this research, to
rapidly respond to customer demands by intelligently
and effectively communicating information on the
factory floor. It aimed to increase the performance of
downstream production instructions flow fed from
parallel upstream flow of information on the factory
state.
3.2 Method
To verify the success of the system, the system
needed to respond rapidly to both the initial scope on
product variations as well as the randomly introduced
new customized product variations. In addition, the
system needed to facilitate the appropriate fixture and
machine configurations and assign program
instruction to the factory floor.
The following method was followed:
1. Use the Central FMCS Control software to
initiate a factory floor scan.
2. Input order’s product configuration into the
system using the graphical user interface.
Randomly input new customized product
specifics.
3. Check the response time, the time between
lodging of order into system and dispatcher of
order program instruction.
4. Repeat steps 2-3, varying the order specifics each
time.
5. Check fixture database updates in the case of a
new product configuration.
3.3 Results
As shown in Table 1, the system managed to
successfully respond to new customized orders,
seamlessly integrate systems and communicate
information on the factory floor. The FMCS had an
80% success rate.
3.4 Discussion of Results
The experiment was successful in identifying the
states of each cell at any given time. Each process,
from the initialisation of the program to when the
assignment of program production control
instructions took place in real time with delays of less
than 3mins. The system was able to obtain real time
responses for the factory states data. The processes
where a new product configuration was introduced
responded within a 2min to 5min response time.
Extra time was needed for the 3D printing of the
fixture. The cellular configured SCMs were able to
communicate within a radius of up to 1,2km
approximately 99% of the time. This process is still
very quick in industrial terms and is far more time and
labour efficient than manual methods. The product
variety was quantified using the equation given
below.



(2)
Where P refers to the different product designs or
types that are produced in the factory.
the number
of distinct products lines, hard product variety and
the number of number of models in a product line,
soft product variety. Subscript j identifies the product
line (Groover, 2008).
Table 2: Results of FMCS Test.
Order
#
Run
#
Identified
Factory
floor state
correctly
Communicated
with Fixture
reconfiguration
system correctly
Communicated
with machine
reconfiguration
system correctly
New
Fixture
Produced
Assigned
Program
instructions
correctly
Time Result
1 1 Yes Yes Yes N/A Yes 8:30 Pass
2 1 Yes Yes Yes Yes Yes 9:45 Pass
3 2 Yes Yes Yes No N/A 11:18 Fail
4 1 Yes Yes Yes N/A Yes 11:30 Pass
5 1 Yes Yes Yes Yes Yes 14:03 Pass
ICINCO 2016 - 13th International Conference on Informatics in Control, Automation and Robotics
150
The graph below shows how the product variety of a
factory overtime using the Factory Middleware
Communication System (FMCS).
Figure 5: Graph of Product variety versus time.
From the graph, we can see that the system increase
the scope of product variety of a system, making it
well equipped to handle volatile changes driven by
the market.
3.4.1 Performance and Scalability of
Middleware Software
Further test were done to measure the performance of
the middleware layer. A fundamental measure of
middleware performance is latency. Latency is the
time it takes for a two-way operation to be invoked
between a client and server and obtain the results of
the operation (Roulet-Dubonnet et al., 2013). When
we run the client and server on the Core i7 machine,
with the client and server running on different
machines and communicating over a network. The
latency is 2,500 messages per second (400µs per
message).
4 CONCLUSIONS
The flexibility of the system means that users have an
almost limitless expandability and engineers can
adapt and upgrade the system’s features and
capabilities to meet immediate and future
requirements. The middleware communication
system allows for flexible control and information
exchange in a heterogeneous factory environment
driven by the dynamic customer needs for production
execution. The performance of downstream
production instructions flow fed from parallel
upstream flow of information on the factory state that
were increased with the use of this system.
REFERENCES
European Factories of the future assosiation EFFRA
(2013). Factories of the future.
Bloem, J.,Van Doorn, M., Duivestien, S.,Excoffier, D.,
Mass, R., and van Ommeren, E.,(2014). The Fourth
Industrial Revolution. In:Sogeti. Vint.sogeti.com/[
Accessed 5 Mar. 2016].
Schwab, K., (2016). The Fourth industrial revolution. In:
World Econimic Forum. http://www.weforum.org/
agenda/2016/01/the-fourth-industrial-revolution-what-
it-means-and-how-to-respond.
Qiao, G., Lu, R., and McLean, C., (2000)”Flexible
Manufacturing System for Mass Customization
Manufacturing”
Walker, A., and Bright, G., (2013) Distributed Control
Synthesis for Manufacturing Systems using Customers’
Decision Behaviour for Mass Customisation.
Salvador, F., Martin, P., and Piller, F., (2009) Cracking the
Code of Mass Customization. In MIT Sloan
Management Review, Massachusetts.
"SCADA". Wikipedia. N.P., 2016. Web. 17 Mar. 2016.
D. Bailey and E. Wright, “Practical SCADA for Industry”,
(2003).
R. J. Robles, M. -k. Choi and T. -h. Kim, “The Taxonomy
of SCADA Communication Protocols”, Proceedings of
the 8th KIIT IT based Convergence Service workshop
& Summer Conference, Mokpo Maritime University
(Mokpo, Korea), ISSN 2005-7334, pp. 23.
M Choi, “Wireless Communications for SCADA Systems
Utilizing Mobile Nodes”, International Journal of
Smart Home Vol. 7, No. 5 (2013), pp. 1-8.
"Manufacturing Execution System". Wikipedia. N.p.,
2016. Web. 17 Mar. 2016.
"Enterprise Resource Planning". Wikipedia. N. P., 2016.
Web. 17 Mar. 2016.
Lesne, A., (2011) “Shannon entropy: a rigorous
mathematical notion at the crossroads between
probability, information theory, dynamical systems and
statistical physics,” Pierre and Marie Curie University,
France.
Buda, A., Schuermann, V., Wollert, J., (2010) “Wireless
Technologies in Factory Automation,” University of
Applied Sciences Bochum Germany.
Roulet-Dubonnet, O., Lund, M., and Skavhaug, A., (2013),
“IceHMS, a Middleware for Distributed Control of
Manufacturing Systems," in Industrial Applications of
Middleware Communication System for Fixture based Flexible Manufacturing Systems
151
Holonic and Multi-Agent Systems, 1 ed Berlin,
Germany: Springer Berlin Heidelberg, pp. 95-105.
Mikell P. Groover. (2008). Automation, Production
Systems and Computer-Integrated manufacturing, New
Jersey: 3rd edition, Pearson International Edition.p. 55.
ICINCO 2016 - 13th International Conference on Informatics in Control, Automation and Robotics
152