BUILDING-UP AN AUTOMATED DATA COLLECTION SYSTEM
Petri Oksa, Jussi Nummela, Mikael Soini, Lauri Sydänheimo and Markku Kivikoski
Tampere University of Technology (TUT), Department of Electronics, Rauma Research Unit, Finland
The need for gathering real-time information in building and industrial
environments has increased in recent years. In the building environment, along
with automatic utility meter reading, other information is increasingly needed e.g.
room temperatures or device conditions. In the industrial environment, for instance
in the paper industry, information on paper reel locations is essential when
developing logistics and supply chain management.
This paper presents a particular system definition of the automated data collection
system and clarifies management issues. The study is based on the automated data
collection system (ADCS). ADCS is an open and advanced platform e.g. for
building control and monitoring systems.
This study also analyzes ADCS data loads, quality of transmission, and error
sensibility in different system parties. By using ADCS infrastructure, the usually
inconvenient attachment and registration of new devices can be solved.
Keywords: Data collection, information, automatization, meter reading
2 BASIS OF THE AUTOMATED DATA
Device-independent and open data collection in COLLECTION
different kind of applications is a very important area
under current consideration. An open architecture This chapter presents a common overview for an
consisting of actuators, sensors, transmission media, automated collection system. The basic building-up
secure data transmission, compact messages, and procedure for an automated data collection system
message types presents a big challenge in designing consists of phases such as new device attachment,
the whole system and needs comprehensive study operations during maintenance and system
before implementation procedure. operations. In the attachment phase, new device
ADCS is a common name for an open data attachment activities are presented, and in the second
collection system, it is planned to be suitable for phase, operations during ADCS maintenance are
applications such as automatic meter reading explained. In the last phase, the functions of system
(AMR/AMS) , RFID-systems , and for operation are clarified.
building automation sensor networks –. In
AMR systems, for instance, the amount of 2.1 Attaching
transferred data is small. Thus, an ADCS offers a
varied choice and the features required to manage the One of the main goals of the device attachment
data transmission. As Figure 1 in chapter 3 will show, procedure is to be “Plug & Play”. As discussed in ,
the advantage of ADCS is in its versatile options for home-networking solutions should be easy to install,
public utilities and their customers. Many of these providing PnP and/or autoconfiguration features, and
features suggest the desirability of real-time should enable remote maintenance from the service,
processing and methodology which is one of the network, or manufacturer site. Reliability and
main motivations behind this paper. robustness are also considered mandatory, as
This paper is organized as follows: chapter 2 residential users will have difficulty identifying and
clarifies the basis of the automated data collection. handling problems and home-networking products
Chapter 3 presents a solution for ADCS data need to operate all day and night long .
collection system on the basis from chapter 2. When an actuating device is installed, the
Chapter 4 presents a case study for ADCS data concentrator automatically updates itself with the
collection and chapter 5 concludes the study and values from the device registers, for example device
takes a look at future work. number- and energy consumption values. The
exception is a pulse reading technique that cannot be
read via the device register.
Ubiquitous Computing and Communication Journal 1
The information is sent to the main system to rarely are these systems based on an open
ensure that a new device is connected to the system. infrastructure.
The same information is supplied to all relevant Automated configuration will simplify the
subsystems. This should minimize the occurrence of system operator's task of building and maintaining
human mistakes in the information flow , , . the sensor network . As in most novel AMR and
Basically, most of today’s AMR consumption ADCS systems, the new device attachment
meter registration is not based on entirely self- procedure is designed with easy management
configurable or automatic updating methods and functions.
these systems are not generally open-based.
Therefore difficulties often arise in the case of 2.3 System Operation
software updates when vendor support is needed to
solve the problems. In addition to this, some From the system user’s point of view the ADCS
software conflicts cannot be solved or can take a is self-configurable after the registration procedure.
long time to be found. Even a vendor can come up The ADCS network is monitored and maintained by
against software conflicts that cannot be solved the DCU. Two-way communication allows network
without the help of external software consultations control and software updates to be controlled by the
. Also the lack of plug & play standards is an main system programming tools. For instance, meter
obvious disadvantage. reading intervals can be configured from a system
A new AMR consumption meter installation and management site, so there is no need for additional
configuration is normally handled by an electrician pre-configuration for the sensing or metering devices.
visiting the building and reconfiguring the main Also all other software updating is handled by the
system. Commonly, the consumption meter is pre- main system applications.
configured by a vendor according to the The necessity for data content is greatly
requirements of a utility company. After a meter is dependent on device-based definitions. Table 1
installed in a building whose energy consumption is outlines the basic information needed to complete the
to be metered, the rules of communication between data transmission. Certainly, much more detailed
the meter and main system are established by the information can be acquired, but at the same time it
means of main system tools. In the last phase, the increases the amount of data transmission , .
metering information (for example meter ID number, The amount of data transfer should be kept as small
energy consumption, type of meter, meter location or as possible.
property information) is stored in the meter value
database. 3 A SOLUTION
In , the biggest challenge in all symmetric
security systems is how to exchange the initial This chapter presents a solution for automated
encryption key safely. In many communication data collection system and clarifies the ADCS
protocols, this is carried out without any security functions. Under subheading A, the data collection
procedure or the challenge has been left to the procedure is clarified. Below subheading B, system
application developer to solve , , and . If operation is presented and below subheading C, the
an adversary receives this first insecure message, the DCU (Data Collection Unit) registration is
security of the whole network is threatened. One determined. Under the subheading D, analysis of the
solution to this problem is to exchange the primary ADCS solution is given.
key through physical contact .
Novel sensor registration should ensure security 3.1 Data Collection
procedures to authenticate new devices safely into
the system with easy and fast registration, for For an ADCS, a metering device could be for
example as executed in . In practice, this means instance a utility meter, actuator or sensor. The
that a new device receives a primary key from the communication between a metering device and a
registering device that is synchronized with the master device is two-way based as seen in Figure 1.
network master node. This means that In this architecture model, the heart for data
communication in the new network is immediately collection is the Master DCU (MDCU). The MDCU
secure . operates as a data server and data storage. All users
can access the data through the MDCU. The Slave
2.2 Operations during Maintenance DCU (SDCU) can also store data but it also forwards
the information to the MDCU. The SDCU includes
An important and essential issue in future database storage in case of data transmission errors
metering systems is the establishment of automatic (packet corruption or loss) or total blackouts. When a
configuration  and easy installation methods. So data transmission blackout or error occurs, the
far there are many equipment suppliers whose metering information is saved in the SDCU database
systems are designed to operate automatically, but and retrieved by the MDCU during the next MDCU
Ubiquitous Computing and Communication Journal 2
reading schedule. Thus, the metering information If the MDCU somehow does not receive the
will not be lost. Further, the MDCU includes a message, then the Type C message is retransmitted
database backup in case of power- and data until MDCU receives the message and sends the
transmission failures. acknowledgement (Ack) message to complete the
registration (message type B, see Table 1 for
message types). A notification message from the new
device attachment is then sent to the SDCU and
stored in the database in the MDCU. This message
consists of at least the new device’s ID number and
name. After this is done, the user who is logged on to
the system can view the stored metering data and
information from the database to exploit, for
example, a meter reading value or sensor data ,
Appropriate message types and descriptions in
ADCS registration are presented in Table 1. The
functionality of each message type is also described.
Table 1: Example message types in ADCS.
Message Description Function
Type A Request DCU DCU attachment to
Type B Acknowledgement Ack to MDCU
Type C Negative Ack Retransmission to
Figure 1: The main parts of ADCS data collection MDCU
Type D Data content Data content from meter/
The ADCS infrastructure is designed for
automatic configuration. The basic procedure is that
all network maintenance is configured by the MDCU. According to Table 1, four message types are
In this way, the SDCU is pre-configured at a new proposed: Type A, Type B, Type C and Type D.
DCU attachment. This means that when a new These short messages do not need large data content
metering device is within range of the MDCU, it and do not substantially increase the data packet
updates all necessary information from the MDCU. length. Data packets include the pre-defined
After that, a new metering device is ready to collect metering information such as energy usage,
the pre-defined metering data, for example, from an temperature, or moisture (see Table 2).
apartment building environment. A new device authentication is carried out by
user authentication when a new user enters his or her
3.2 DCU Registration username and password. After logging on to the
system the user can view the metering data. Because
One of the most common architectures in data of minor secrecy demands and the nature of the
collection systems is a centralized architecture. In collected data, a primary key exchange is not needed.
this architecture, the MDCU manages registration,
authentication, and device control and monitoring. 3.3 System Operation
Other general types are semi-distributed and
distributed architectures, presented in . In this The SDCU is pre-configured by the MDCU
paper, centralized architecture is selected because of software tools before the attaching procedure. In all
its suitability for small systems where the amount of such procedures, the MDCU recognizes the device,
collected data is also relatively small . executes the registration with the new device,
First, the attaching of a new metering device to updates the metering database, and maintains
the ADCS system is defined. When a new DCU network operation. The network operation includes
comes within the ADCS identification range, the control and maintenance operations, operations
MDCU wakes up and automatically sends a query under maintenance, and the metering interval
message through the network to the new metering changes and configurations.
device. The other MDCU or SDCU works as a When the aggregate of sent and received queries
repeater to forward the query message to the from the database is estimated, the data load can be
destination MDCU. After that the MDCU verifies checked with ADCS centralized architecture . To
the message format and if the format is correct, the simplify, the more queries made the more data
MDCU confirms that the message includes valid transmission and load. Moreover lots of queries lead
content for the data exchange. to larger network requirements where the main focus
Ubiquitous Computing and Communication Journal 3
is to enable sufficient data transmission, and to avoid load profile, system functionality and usability
congestion problems. So, this assumes that data estimation is presented . In the system
packet size should be minimized or a network must application layer the ADCS is data load effective,
have sufficient capacity to carry out the data self-configurable and useful in the building
transmission demands. The ADCS also supports automation environment where changes in the
integrated push- and pull based queries over the environment are commonly predictable. Also the
hierarchy. authentication can be executed without primary key
In ADCS architecture, data load is optimized by exchange. Short messages and small data packets are
cutting query requests and message lengths to a the starting point in defining of ADCS data
minimum. Hence, the query sequence is in such a transmission. The number of messages is also small;
form as to be suitable for any kind of data collection only the required data is transmitted.
network type and is also self-configurable for The basis for more accurate designing and
network and device extensions. This paper discusses specifying an application-specific automated data
data load and network requirements only generated collection system is presented in .
in the ADCS system itself.
Table 2 compares different metering 4 VERIFICATION: A CASE STUDY
characteristics and suitable environments for ACDS.
Some less obvious differences are in need of This chapter presents a case study for the
different metering parameters between domestic and proposed ADCS data collection. The case study
industrial building environments. exploits a centralized network architecture model
presented in . The network architecture model is
Table 2: Examples and comparison of different evaluated by a simulations procedure to verify the
metering characteristics and suitable metering overall performance in different traffic conditions.
Value Unit A B 4.1 Data Collection- and Traffic Simulations
Electricity kWh x x
Quality control e.g. outages x x A method for system verification is determined
Temperature °C/K x x under this subheading. First, the system architecture
Moisture ppmv x x is presented and then verified with the simulation
Data load kbs/Mbs x x
Humidity % x x
procedure presented under subheading 4.2. This
Lighting conditions lx x x simulation procedure concentrates on building
HVAC °C/K x x automation activities. Figure 2 presents ADCS data
A=domestic building environment (apartment buildings, single- collection features in a basic building automation
family houses, rented flats) environment.
B=industrial building environment
Because of a small amount of transferred data,
possible retransmission does not generate notable
external traffic load on the transmission network.
The ADCS also implies the requirement of
supporting multihop networking and the flexibility of
data transmission that can be used to quickly modify
or exploit the infrastructure. A significant feature of
ADCS network is the detection of data transmission
failures that improve the quality of transmission.
These networks are also inherently self-healing, so
users do not have to worry about losing
communication with control devices across the
building automation system .
Combined with wired- and wireless based
transmission media, the ADCS offers a good solution
for sparsely populated areas. Especially in novel
meter reading systems, which support a wireless
infrastructure, easy device installation and
attachment, these multiform transmission media
hybrids are extremely valuable.
In this chapter, the analysis of the ADCS data Figure 2: The ADCS data collection features
Ubiquitous Computing and Communication Journal 4
4.2 Simulation Results wireless 433 MHz is another issue to be solved in the
future. This study does not commit itself to a specific
To evaluate traffic load, the centralized network structures or parameters, and therefore,
architecture model and its usability for ADCS in these appear likely to be the next major focus for
practice, the information flow simulation was future research projects.
executed. The centralized data collection architecture
of ADCS is shown in Figure 3. ACKNOWLEDGEMENT
The authors would like to thank the National
Technology Agency of Finland (TEKES) for funding
 M. Kezunovic, and T. Popovic, “Substation
Data Integration for Automated Data Analysis
Systems,” in IEEE Power Engineering Society
General Meeting, pp. 1–6, June 2007.
 R. Panos, and T. Freed, “The Benefits of
Automatic Data Collection in the Fresh Produce
Supply Chain,” in Proceedings of the 3rd
Annual IEEE Conference on Automation
Science and Engineering, Scottsdale, AZ, USA,
Figure 3: The ADCS network architecture pp. 1034–1038, September 2007.
 M. Soini, “Studies towards Developing
The centralized architecture model for Wireless Control and Monitoring Reliability
simulation consists of one MDCU that works as a and Usability in a Building Environment,” Ph.D.
server, and DCUs that work as data storage and thesis, Dept. of Electrical Engineering, Tampere
forwarders. This centralized architecture model is University of Technology (TUT), Tampere,
comprehensively studied in . In accordance of 2007.
this study, the centralized architecture is suitable for  F. Zhao, “Wireless Sensor Networks: An
ADCS where traffic loads are high, maintenance is Information Processing Approach,” Elsevier
remarkably easy, modifiability is very good and the Inc., 2004.
number of DCUs is small.  G. Baweja, and B. Ouyang, “Data acquisition
Table 3 introduces some common selection approach for real-time equipment monitoring
parameters for ADCS establishment. By using varied and control,” in IEEE/SEMI Advanced
parameters and case study results, utility personnel Semiconductor Manufacturing Conference, pp.
can achieve benefits for data transmission 223–227, May 2002.
development of an automatic data collection system.  Q. Han, I. Lazaridis, S. Mehrotra, and N.
Venkatasubramanian, “Sensor Data Collection
5 CONCLUSION AND FUTURE WORK with Expected Reliability Guarantees,” in
Proceedings of the 3rd IEEE Int’l Conf. on
Accurate and reliable data collection is an Pervasive Computing and Communications
important subject for the building environment and Workshops (PerCom 2005 Workshops), pp.
its development issues. Another big challenge is to 374–378, March 2005.
form an open infrastructure which can be used in  T. B. Zahariadis, “Home Networking
most buildings to meter and control different Technologies and Standards,” Artech House
attributes of changes in living conditions. Inc., 2003.
This paper presented an open architecture to  Concept of Large AMR, Oy Comsel System Ab,
attain diversified control and a monitoring platform 25 Pages, 2004.
especially for the building environment. The  Enermet Metering Solutions, Enermet Oy, 2005.
platform can consist of a sensor, utility meter or Available:
other actuator. New device registration and http://www.enermet.com/en/metering/
attachment to a communication network is also  Guidance for Metering in Federal Buildings,
defined. One of the biggest motivations of this study Continental Automated Buildings Association
is to solve the problem of new device registration (CABA), 22 Pages, February 2006.
and attachment.  H. Sikkilä, M. Soini, P. Oksa, L. Sydänheimo,
To make ADCS sufficiently versatile for and M. Kivikoski “KILAVI Wireless
common wired-based applications such as TCP/IP or Communication Protocol for the Building
Ubiquitous Computing and Communication Journal 5
Environment – Security Issues,” in Proceedings the 33nd conference on Winter simulation 2001,
of the 10th IEEE International Symposium of Arlington, Virginia, pp. 984–990, December
Consumer Electronics (ISCE2006), St. 2001.
Petersburg, Russia, pp. 640–645, June 2006.  W.C. Cheng, Cheng-Fu Chou, L. Golubchik, S.
 A. Perrig, R. Szewczyk, V. Wen, D. Culler, and Khuller, and Yung-Chun Wan, “A Coordinated
J. D. Tygar, “SPINS: Security Protocols for Data Collection Approach: Design, Evaluation,
Sensor Networks,” in 7th Annual ACM and Comparison,” in IEEE Journal on Selected
International Conference on Mobile Computing Areas in Communications, Vol. 22, Issue 10, pp.
and Networks (MobiCom 2001), pp. 189–199, 2004–2018, December 2004.
Rome, Italy, July 2001.  M. Soini, H. Sikkilä, P. Oksa, L. Sydänheimo,
 C. Karlof, N. Sastry, and D. Wagner, “TinySec: and M. Kivikoski, “KILAVI Wireless
A Link Layer Security Architecture for Communication Protocol for the Building
Wireless Sensor Networks,” in Proceedings of Environment – Networking Issues,” in
the 2nd ACM Conference on Embedded Proceedings of the 10th IEEE International
Networked Sensor Systems, pp. 162–175, Symposium of Consumer Electronics
November 2004. (ISCE2006), St. Petersburg, Russia, pp. 453–
 J. Undercoffer, S. Avancha, A. Joshi, and J. 458, June 2006.
Pinkston, “Security for Sensor Networks,” 2002.  F. J. Groen, Jiang Siyuan, A. Mosleh, E. L.
 S. D'Antonio, M. D'Arienzo, A. Pescape, and G. Droguett, “Reliability data collection and
Ventre, “An Architecture for Automatic analysis system,” in Annual Reliability and
Configuration of Integrated Networks,” in Maintainability Symposium (RAMS 2004), pp.
IEEE/IFIP Network Operations and 43–48, 2004.
Management Symposium (NOMS 2004), Vol.  Siemens Building Technologies Pioneers
1, pp. 351–364, April 2004. Wireless Solution for St. Anthony's Sensitive
 K. L Headley, D. Davis, D. Edgington, L. Environment, Available:
McBride, T.C. O'Reilly, and M. Risi, http://sev.prnewswire.com/computer-
“Managing Sensor Network Configuration and electronics/20070927/NYTH14627092007-
Metadata in Ocean Observatories Using 1.html.
Instrument Pucks,” in The 3rd International  J. Nummela, P. Oksa, L. Ukkonen, L.
Workshop on Scientific Use of Submarine Sydänheimo, and M. Kivikoski, “Comparison
Cables and Related Technologies, pp. 67–70, of Different Information Flow Architectures in
June 2003. Automated Data Collection Systems,” in
 K. Mase, and H. Yamamoto, “Advanced traffic Proceedings of the 5th International Conference
control methods for network management,” in on Informatics in Control, Automation and
IEEE Communications Magazine, Vol. 28, Robotics 2008 (ICINCO 2008), Funchal,
Issue 10, pp. 82–88, October 1990. Madeira, Portugal, May 2008., to be published.
 N. H. Robertson, and T. Perera, “Feasibility for
Automatic Data Collection,” in Proceedings of
Table 3: Some common directive system selection parameters for ADCS
Parameter PLC RF WLAN WiMAX TCP/IP Optical fiber WSN
Max. 100-150 m ~50 m (in ~200-500 m Up to 10 km Up to 100 m ~500 m Up to 20 km 2 km out-
communication building) door-~10 m
Suitability for Average High Average High High Average High High
Data communi- Modem Base station Base station Base station Subscriber Ethernet Line modem Master node
cation interface line
Network Power line Radio IEEE 802.11 IEEE 802.16 Protocol- IEEE 802.3i IEEE 802.3j 802.11,
communication wiring frequency based Bluetooth,
Ubiquitous Computing and Communication Journal 6