PROFIBUS International Conference Coombe Abbey, Warwickshire, UK, June 2005
Paper C1 Page 1 of 7 K. Koumpis et al
Konstantinos Koumpis, Lesley Hanna Mats Andersson and Magnus Johansson
Sira Technology Ltd. connectBlue AB
South Hill, Chislehurst, Kent, BR7 5EH, UK Norra Vallgatan 64 3V, Malmö, SE-21122, Sweden
{costis.koumpis,lesley.hanna}@sira.co.uk {mats.andersson,magnus.johansson}@connectblue.se
Abstract
This paper is about networking of embedded devices
such as actuators and sensors to perform industrial
control and monitoring. Architectures for various types
of applications and technical options for implementing
these architectures are discussed. A review of the state
of the art in cable replacement with an emphasis on
practical examples is also featured. The paper continues
with a hybrid wired/wireless fieldbus topology as well
as the approach of mesh networking in which wireless
nodes function as senders, receivers and routers
allowing distributed control and sensing. Finally, a
technology roadmap for wireless industrial control and
monitoring developed as part of the RUNES project is
presented.
Introduction
An industrial facility typically comprises a relatively
small control room, surrounded by a relatively large
physical plant. The control room is equipped with
panels that depict the state of the plant as captured by
sensors and input devices that control the actuators,
affecting the state of the plant. The actuators and
sensors are often relatively inexpensive when compared
with the cost of the cable that needs to be used to
connect them. The difference becomes even greater
when considering the high installation and maintenance
costs, the high failure rate of connectors and the
difficulty of troubleshooting them.
The information being communicated in industrial
environments is typically state information and as such
in normal operation it takes the form of repeated
streams of small packets. At the same time, these
packets are associated with critical tasks having strict
timing requirements in harsh environments. The latter
may include extremely high or low temperatures, high
humidity levels, intense vibrations, explosive
atmospheres, corrosive chemicals and excessive
electromagnetic noise caused by large motors and
conductors. Thus, in general, the required data
throughput of the network is relatively low, but its
reliability needs to be very high.
In recent years, the desire for connectivity and physical
mobility has caused an exponential growth in wireless
communication systems. Wireless telephony has
entered our daily lives and wireless local area
networks increasingly serve as a means to access
business and private data. In industrial environments,
apart from lower installation and maintenance costs,
wireless systems can offer ease of equipment
upgrading and practical deployment of mobile robotic
systems and micro-electromechanical systems
(MEMS).
The rest of this paper is organised as follows. The
next two sections provide an overview of the
evolution in wired control and monitoring and present
solutions in cable replacement along with examples
of successful deployments. The paper continues with
the hybrid wired/wireless fieldbus and mesh
networking approaches. Finally, the results of study
on how wireless technology is expected to impact
industrial control and monitoring are presented in the
form of a technology roadmap.
Wired Control and Monitoring
One may identify three main paradigms in wired
communication for industrial control and monitoring:
a) parallel wiring, b) fieldbusses and c) industrial
Ethernet. According to the first paradigm, each of the
field devices is connected with parallel wires to the
I/O module of a control or monitoring system. Such a
point-to-point wiring approach became obsolete
following the introduction of fieldbus technology,
which allowed the use of only one two-wire line to
provide power, control and configuration functions to
devices. Fieldbusses allow many devices to be
connected on the same wire and provide the
necessary addressing mechanism to support
communication with them. The open standardisation
approach adopted by main fieldbus technologies such
as Profibus has facilitated interoperability among
systems from multiple vendors and has been proven
in many factory, process and building automation
applications worldwide.
Several fieldbus manufacturers have more recently
recognised the advantages of Ethernet, which is the
established standard bus system in the office world,
for industrial applications [7]. The advantages are
related to the physical layer, particularly in terms of
PROFIBUS International Conference Coombe Abbey, Warwickshire, UK, June 2005
Paper C1 Page 2 of 7 K. Koumpis et al
bandwidth, which can be higher than 100Mb/s rather
than up to 12Mb/s for fieldbusses. The higher
bandwidth can be utilised by larger packets e.g., for
computer vision systems. In the past, the main
disadvantage of Ethernet was the Carrier Sense Multiple
Access / Collision Detection (CSMA/CD) contention
protocol which does not guarantee time-critical
communication. However, by splitting up the network
in multiple collision domains using switches (or
bridges), the prices of which have dropped dramatically
as a result of the Internet revolution, every port on a
switch is in its own collision domain and as such no
more collisions between devices attached to the switch
take place. Ethernet-based solutions offer also improved
information sharing between manufacturing and backoffice
systems such as asset management and inventory
control applications. The data can also be made easily
available via a Web server for remote control and
monitoring purposes.
Cable Replacement
Cable replacement is used here to denote the
elimination of wires as the physical layer to carry data
without requiring any physical changes to the
machinery/instrument, the control panel or the
underlying software involved. Industrial devices using
traditional serial interfaces such as RS232, RS422 and
RS485 are good candidates for cable replacement. This
is because serial interfaces are typically connected to
standard PCs and the connecting software is application
dependent or device specific. Apart from serial point-topoint
connections, cable replacement has found
applications into master/slave multi-point connections,
wireless parameterisation and diagnosis particularly to
do with moving/rotating subsystems, e.g., robotic arms.
There are limitations to what cables can be replaced,
though, due to the error characteristics of wireless links.
When deterministic communication with latency (the
time it takes for a packet of data to get from one
designated point to another) under 10ms is essential,
wireless transmission should be avoided.
Wireless transmission can take place in various
frequency bands and the transmission power is often
restricted by law. The 2.4GHz Industrial Scientific and
Medical (ISM) band is the most widely used. The
900MHz band, which is characterised by lower
throughput but better range and wall penetration, is only
available in a few countries and used by proprietary
protocols. The 5.8GHz band holds a lot of potential in
terms of higher throughput, better noise immunity and
smaller antennas, but products are yet to be proved in
the market.
Table 1 compares three main wireless transmission
technologies. These technologies are complementary
rather than competing to each other, as they address
different needs and have different strengths.1
Bluetooth [2] requires a low-cost transceiver chip in
each device to be connected. Each device has a
unique 48-bit address and the transceiver transmits
and receives in the ISM band. Connections can be
point-to-point or multipoint with a range of 20-100m.
Data can be exchanged at a rate of 1-3Mbps and a
Frequency Hopping Spread Spectrum (FHSS) scheme
allows devices to communicate even in areas with a
great deal of electromagnetic interference. However,
this makes it extremely difficult to create extended
networks without large synchronization cost. Built-in
encryption and simple verification is also provided by
Bluetooth.
ZigBee [8] moves data only a quarter as fast as
Bluetooth but can handle orders of magnitude more
devices at once and has been optimized for low
power consumption. This low power consumption is
achieved by the Direct Sequence Spread Spectrum
(DSSS) which allows devices to sleep without the
requirement for close synchronization. Another
spread spectrum technique under development, Ultra-
Wide Band (UWB), broadcasts simultaneously on a
very large frequency range at low power. The idea is
that the signal is spread so thinly that interference
will be negligible in any given frequency. UWB is
expected to be able to deliver high throughput,
particularly in areas with physical obstacles.
Standard
(market name)2
802.15.1
(Bluetooth)
802.11b
(Wi-Fi)
802.15.4
(ZigBee)
Application
focus
Cable
replacement
Web, email,
video
Control &
monitoring
Bandwidth
(Kbps)
1000-3000 11000 20-250
Transmission
range (m)
20 (Class 2)
100+ (Class 1)
100+ 20-70, 100+ (ext
amplifier)
Nodes supported 7 32 264
Battery life
(days)
1-7 .5-5 100-1000+
Power
consumption
(transmitting)
45mA (Class 2)
<150mA (Class
1)
300 mA 30 mA
Suitability for
low duty-cycle
applications
Poor (Slow
connection time)
Poor (Slow
connection time)
Good
Spread spectrum
technology
FHSS DSSS DSSS
Memory
footprint (KB)
50+ 70+ 40
Success metrics Cost,
convenience
Speed,
flexibility
Power, cost
Table 1 A comparison of major wireless standards
using the ISM band.
1 Infrared (IrDA) can also be considered for cable
replacement but has the shortcoming of requiring line of
sight which limits its applicability.
2 The IEEE 802 standards typically create the specifications
at the physical layer and portions of the data link layer. The
higher layer protocols are left to the industry and the
individual applications. Hence the standard and market
names are not always interchangeable.
PROFIBUS International Conference Coombe Abbey, Warwickshire, UK, June 2005
Paper C1 Page 3 of 7 K. Koumpis et al
Some examples of industrial applications where cable
replacement has been successfully deployed are given
below.
Phoenix Contact (Germany) use Bluetooth to replace
parallel wiring. A wireless multiplexer can replace
wires for up to 32 digital inputs, 32 digital outputs, 2
analogue inputs and 2 analogue outputs. The product
mirrors the inputs to the outputs and vice versa. The
latency from input to output is 10 ms or less.
Expert Monitoring (UK) have developed WiSNet which
allows sensors or instruments to be installed wirelessly.
A typical WiSNet system consists of an Ethernet / USB
network controller and sensor transmitter modules that
enable sensors to be positioned anywhere inside or
outside manufacturing plants. The transmitters use
Bluetooth to send sensor data to a wireless network
controller unit connected to a PC.
EnVision (USA) have developed a wireless sensor that
enables bioprocessors to monitor fermentation and cell
culture processes directly in reactors. The sensor can be
configured by either a Bluetooth wireless browser-based
user interface or configured for Foundation Fieldbus
communications.
Bromma Conquip (Sweden) use Bluetooth to replace
the cables connecting the control systems of harbour
conveyor cranes with configuration and maintenance
tools.
Schneider Toshiba Inverter Europe (STIE) (France) use
Bluetooth to replace the cables that connect a
configuration tool (running on a PC or PDA) to their
family of inverter products.
Hybrid Wired/Wireless Fieldbus Networks
The need to retrofit any support for wireless to existing
installations instead of creating new systems from
scratch, increases the requirement for hybrid
wired/wireless fieldbus networks. The standard Profibus
Decentralised Periphery (DP) network is based on a
standard RS485 interface but often using unusual baud
rates (93.75 and 187Kbps). The protocol has timing
requirements; however, these requirements are
adjustable because they were originally introduced to
support modems. In order to be able to replace a
fieldbus such as Profibus DP with a wireless link, it is
very important to keep messages unbroken to support
the timing requirements of the fieldbus protocol.
Figure 1 depicts three fieldbusses that incorporate fixed
wireless devices. At the low baud rate of 9.6Kbps the
full number of 7 wireless nodes in a Bluetooth network
can be supported, however at 93.75Kbps only 2-3 nodes
can be supported. The maximum bandwidth in a pointto-
point configuration is 187Kbps. Detailed studies on
hybrid Profibus architecture supporting mobility can be
found in [1]. The support of industrial multimedia
traffic over industrial wireless fieldbusses has also been
studied [5].
Figure 1 Example of a hybrid wired/wireless profibus
topology.
Modbus Remote Terminal Unit (RTU) fieldbus also
uses RS485 as a transport media and has been
successfully replaced by a Bluetooth wireless link
both in single-point and multi-point configurations.
Examples include the Phoenix Contact and STIE
products mentioned above which use Modbus RTU
on top of wireless links.
Mesh Networks
The traditional wireless systems have mostly used
base station style radio links, with point-to-point or
point-to-multipoint transmission. These wireless
approaches have liabilities in industrial applications
such as rigid structure, meticulous planning
requirements and dropped signals. Rather than
relying on a central communication coordinator and
its associated reliability and efficiency issues, it is
possible to use a collection of wireless devices
maintaining connectivity to create a path for data
packets to travel. This approach is known as a mesh
network (Figure 2) [3] and in many ways it resembles
an idealized version of a top-level Internet backbone
in which physical location is less important than
capacity and network topology. In mesh networks
each node has a low-power transmitter and
communicates directly only with neighbouring nodes
and these latter nodes relay data to more distant
nodes. Should a link become congested or a node fail,
the mesh automatically redirects data packets via an
alternative path. This characteristic makes mesh
networks virtually immune to localised interference
such as that caused by motors turning on or arc
welders.
Mesh networks allow applications to grow based on
demand (the addition of new nodes is relatively
simple), limiting fixed costs and providing great
flexibility and capacity. They also minimise the need
for elaborate site surveys or physical modifications to
plants. Moreover, because of the short range of each
transmission, the approach provides better utilisation
PROFIBUS International Conference Coombe Abbey, Warwickshire, UK, June 2005
Paper C1 Page 4 of 7 K. Koumpis et al
of available bandwidth than systems using high-power
transmitters. Unlike other approaches, in mesh networks
the major design requirement is the lowest possible
node cost.
control/monitoring station
Figure 2 Wireless mesh network topology.
Many important questions remain though with respect
to power management, routing strategies and algorithms
in mesh networks. Even if good answers were to be
given, mesh networks would not displace all existing
industrial networks since deterministic operating modes
(low latency and jitter) cannot be guaranteed.
Technology Roadmap
A technology roadmap for industrial control and
automation was developed in the context of the RUNES
project. The goal of technology roadmapping is to plot
the future development of a technical field and help set
more competitive and realistic targets. The objective of
RUNES is to derive architectures and provide
middleware and specialised simulation and verification
tools that enable the creation of large-scale, widely
distributed, heterogeneous networked embedded
systems which interoperate and adapt to their
environments. The roadmap is composed of the
knowledge and views of over 25 organisations (large
and small companies and research institutes) collected
across Europe between Oct 2004 and Apr 2005 [6]. The
focus of the roadmap is approximately 10 years
although timescales for technological progress are
notoriously difficult to predict. Below we summarise
the technical, organisational and social issues identified.
Action plans for strategic positioning and resource
allocation are also defined. Apart from the textual
description of the roadmaps, we have summarised some
of the findings in graphical form in Figure 3. The span
of different technologies does not necessarily coincide
with the first specification releases but mainly with the
time adoption started. Regarding the decline of
particular technologies, some assumptions, based on the
information currently available, are made.
Technical Issues
Security
Major concerns about the integrity of signal
transmission and reception have been expressed by
industrial end-users who are worried about leaving
their control and business systems vulnerable to
hacking or denial of service attacks. Several solutions
are based on proprietary protocols designed
specifically with security concerns in mind, although
this can be a barrier for future upgrades with standard
equipment. Spread spectrum schemes provide
inherent security against jamming or interference.
Systems can employ 128-bit encryption with
dynamically generated keys in order to prevent
eavesdropping and unauthorised access. However,
this is computationally too expensive for
deterministic communication. Continuous monitoring
of network activity and attempted access should also
be supported. Spread spectrum technology provides
inherent security against jamming or interference.
Robustness
The consequences of unreliable control and
monitoring are not trivial. Industrial applications
entail the risk of substantial losses through equipment
damage, personnel injuries, loss of raw materials and
environmental pollution. From the perspective of the
integrator or end user, communications expertise and
comprehensive technical support are key
considerations that distinguish ordinary wireless
products from the robust wireless communication
solutions required by industrial users. Many of
today’s wireless solutions offer mean time between
failures which are unacceptable to deliver sustained
performance in harsh environments. Any early
failures could slow adoption of wireless.
Encapsulation of sensitive electronic components is
particularly important to achieve improved product
reliability and extended lifecycles in harsh
environments. Positioning of antennas is also critical
in order to ensure reliable operation, as it can affect
the bandwidth and data transfer rates.
Fail-safe/fail-soft operation
Designers of wireless industrial systems need to
provide redundancy and allow degradation. Systems
must be designed to go into a safe mode if and when
a failure happens (fail-safe). Systems must also be
capable of operating at a reduced level of efficiency
after the failure of a component or power source (failsoft).
Placing devices on different networks provides
substantial fail-soft operation in that the failure of a
single network only affects the devices that it
supported. Devices connected to other, unrelated and
perhaps geographically distributed, networks would
be unaffected by the events that caused the failure of
the first network.
PROFIBUS International Conference Coombe Abbey, Warwickshire, UK, June 2005
Paper C1 Page 5 of 7 K. Koumpis et al
Interference immunity
Multi-path fading is caused when multiple copies of a
source signal arrive at a receiver through different
reflected paths and affects wireless communications
indoors. The phase variance in the signal copies can
result in interference that reduces signal strength,
effective range, and data transfer rates. A wireless node
has to handle multi-path signals, but interference can be
caused by signals from other nearby systems
particularly for unlicensed frequency bands. Extremely
critical wireless equipment can operate inside a Faraday
cage.
Power availability
As the speed of embedded processors increases and
more peripherals are integrated into a single chip, the
applications that run on these devices become more
computationally intensive. However, technological
advances of the batteries which power the embedded
systems lags significantly behind, and as a result, power
consumption is one of the most important issues for
mobile wireless embedded systems. The major sources
of energy waste include packet collisions, overhearing
unwanted packets, control packet overheads and idle
listening. For monitoring applications unnecessary high
sampling rates also waste energy. One answer to the
power problem is to scavenge power from the industrial
environment. For instance, this could include
development of photoelectric cells that draw energy
from lighting. Power scavenging technology is
something that industrial sector are thinking about
already as supplying power to remote areas is costly.
Interoperability
At present, there are a number of different protocols for
fieldbusses and industrial Ethernet, and end-users are
concerned about the long-term cost implications of
installing closed wireless systems. Until standards have
been established and the market has become one for
volume suppliers, fears of being left with an obsolete
system may hold back widespread adoption. An
important component towards interoperability is
middleware, which defines appropriate abstractions and
mechanisms for dealing with the heterogeneity of
devices. Because ideally devices must operate
unattended, the middleware has to provide new levels of
support for automatic configuration and error handling.
Development of middleware is a key activity within the
RUNES project [4]. The IEEE Smart Transducer
Interface for Mixed-Mode Communication Protocol
(1451.4) was also revived last year in recognition of a
need for a standardised approach to device interfaces.
Interfaces
Most industrial processes now modelled and controlled
in IT systems, but they require the user to be physically
at a terminal. Given the increasing need for workers to
access systems remotely more research on machine to
human interfaces needs to be undertaken. Human to
human interfaces via devices that allow voice, image
and video communication is another potentially
important field given that they provide natural and
real-time exchange of information, e.g. when
someone on a factory floor needs to report a fault to
an expert. Very few projects have addressed
multimodal interaction such as gesture based
programming of robotic arms.
Organisational Issues
Culture
The industrial automation sector, in general, is very
conservative. Companies do not want to take chances
with large investments in new installations and
require demonstration of practicality (preferably by
someone else). People involved in risk management
are not receptive to new technologies. However,
cheaper, faster, safer and more reliable options are
always there for successful innovators. Often,
employees who misunderstand new technology and
lack confidence in its ability to improve over
previous practice use many new systems in a
suboptimal way. Collaboration between control
people and radio people is not close enough. Major
cultural differences exist between Europe and the less
cellular-oriented US.
Work force
Industrial companies are expected to face a strategic
human resource issue. The shift from a wired to a
wireless plant, particularly one that can operate
autonomously, will require adjustments in the work
force. Engineers should understand radio
technologies and be able to handle false alarms;
hence revised technical training will be required. A
shortage in embedded software development skills
has also been identified.
Social Issues
Workforce
In general new levels of automation devalue unskilled
labour through its replacement with less expensive
systems. This is expected to increase job security
concerns for people who only posses skills in
declining technologies.
Environmental
Industrial control and automation have an enormous
direct and indirect impact on the environment. New
stricter regulations have been introduced and this
trend is expected to continue. The deployment of
large numbers of wireless sensors can be used for
instance in warning systems to reduce the risk of
environmental pollution.
PROFIBUS International Conference Coombe Abbey, Warwickshire, UK, June 2005
Paper C1 Page 6 of 7 K. Koumpis et al
2004 Vision
Triggers
Targets
Issues 2006 2008 2010 2012 2014
Current levels
Wireless networks open up opportunities for higher flexibility, cost advantages for the installation and operation of
industrial systems
Reduce installation cost by
50% and running cost by 30%
Wireless is the
default choice in
industrial control
and automation
Interoperability
with legacy
systems
Reconfigurable
devices
Scavenging
power from the
environment
Unauthorised
network access
impossible
RUNES Technology Roadmap for Industrial Control and Automation
Fieldbus
IEEE 802.11 (Wi-Fi)
IEEE 802.15.14 (ZigBee)
Conventional batteries
IEEE 802.15.1 (Bluetooth)
IPv6 and Mobile IPv6
IEEE 1451
RFID
Wireless Industrial Ethernet
Middleware
Industrial Ethernet
Collaborative R&D
RF regulations
Environmental regulations
Retrain employees
Power scavenging
actions
social
organisational
technical
Skills shortage
Focus on wireless sensing
Focus on wireless control
IrDA
MEMS
Ultra-Wide band
New materials for encapsulation
Culture change
Reduce installation cost by
80% and running cost by 50%
Figure 3 The RUNES technology roadmap for industrial control and automation in graphical form.
Actions
The use of wireless systems for industrial applications is
in its infancy. The adoption period is expected to be
longer than other sectors (building automation and
control, medical care, disaster response and automatic
meter reading) reviewed in the RUNES technology
roadmaps as end-users migrate incrementally from wire
to wireless. Companies dealing with automotive, food
processing, petrochemical and asset tracking
applications were identified as the early adopters.
A clear difference in the adoption time scales between
wireless in control and monitoring was revealed. While
technologies are maturing, wireless will not be used for
critical control applications. Monitoring in hazardous
and inaccessible areas will be given priority in the
short/medium term and in moving towards this some
lessons can be learnt from successful automatic meter
reading deployments. Many wireless systems on the
market today do not meet local/national regulations,
because they transmit too much power or operate in
frequencies that are not approved for unlicensed use.
Therefore, it is important to determine whether or not
the radio subsystems can be programmed to meet these
regulations. Since 2003 the ATEX directive has become
mandatory for all electrical and mechanical equipment
used in potentially explosive atmospheres and any new
networked embedded components will need to
comply with it.
The required R&D will involve expertise and
engineering skills in communications, sensors and
industrial computer systems which are very difficult
to find under one corporate roof. Partnerships will
therefore be essential. Sources for funding and
mechanisms for facilitating collaborative R&D must
be identified. Industrial companies collaborating with
academia should persuade the latter to reconnect
software and hardware education in their curricula.
Concluding Remarks
The traditional approach to the design of industrial
control and monitoring systems – designing an
architecture that integrates actuators and sensors
within a single physical platform under centralised
control – is changing. The emerging view, as
motivated by new large-scale applications in the
factory floor is one in which actuators, sensors,
computing, and human interfaces may be distributed
across multiple physical platforms. Adopting this new
view of industrial systems design requires researchers
to address a range of issues, architectures and
applications related to wireless connectivity. In
industrial control and monitoring there are certainly
many future alternatives, however, the process of
PROFIBUS International Conference Coombe Abbey, Warwickshire, UK, June 2005
Paper C1 Page 7 of 7 K. Koumpis et al
technology roadmapping helps narrow down the field of
possible solutions to those more likely to be pursued.
Most of the issues and technical requirements are not
orthogonal and as such trade-offs depend on
applications.
Acknowledgements
This work was undertaken as part of the RUNES project
which is funded by the European Commission (contract
IST-004536).
References
[1] M. Alves, E. Tovar, F. Vasques, K. Roether and G.
Hammer; “Real-time communications over hybrid
wired/Wireless PROFIBUS-based networks”, Proc.
14th Euromicro Conference on Real-Time Systems,
pp. 142-151, Jun. 2002.
[2] M. Andersson; “Bluetooth for Industry”, The
industrial Ethernet Book, pp. 10-12, Sept. 2002.
[3] C.-Y. Chong and S. P. Kumar; “Sensor networks:
evolution, opportunities, and challenges”, Proc.
IEEE, pp 1247-1256, Aug. 2003.
[4] P. Costa, G. Coulson, C. Mascolo, G. P. Picco and
S. Zachariadis; “The RUNES Middleware: A
reconfigurable component-based approach to
networked embedded systems”, Proc. 6th Annual
International Symposium on Personal Indoor and
Mobile Radio Communications (PIMRC05),
(Submitted), 2005
[5] J. Haehniche and L. Rauchhaupt; “Radio
communication in automation systems: the Rfieldbus
approach”, Proc. IEEE Int. Workshop on
Factory Communication Systems, pp. 319-326,
Sept. 2000.
[6] K. Koumpis and L. Hanna; “RUNES Technology
Roadmaps”, Deliverable 8.1.2, May 2005.
(available from www.ist-runes.org)
[7] P. S. Marshall; “Industrial Ethernet”, ISA Press,
(ISBN 1-55617-869-7), 2002.
[8] ZigBee Alliance Website,
ليست هناك تعليقات:
إرسال تعليق