30 Published by the IEEE Computer Society 1089-7801/10/$26.00 © 2010 IEEE IEEE INTERNET COMPUTING
Internet of Things Track
E di to r s : Fr é dé r ic Th i es s e • fr e de ri c .t h ie s se @uni sg.c h
Fl or i a n M i ch a h e l le s • f mi c ha he ll es@e t hz .c h
T
he term Internet of Things
1
has
recently become popular to
emphasize the vision of a global
infrastructure of networked physical
objects. Although this vision is com-
pelling, no consensus exists about how
to realize it. The Internet of Things is
partly inspired by the success of RFID
technology, which is now widely used
for tracking objects, people, and ani-
mals. RFID system architecture is
marked by a sharp dichotomy of simple
RFID tags and an extensive infrastruc-
ture of networked RFID readers. This
approach optimally supports tracking
physical objects within well-dened
connes (such as warehouses) but lim-
its the sensing capabilities and deploy-
ment exibility that more challenging
application scenarios require.
We’re working toward an alterna-
tive architectural model for the Inter-
net of Things
1
as a loosely coupled,
decentralized system of smart objects
— that is, autonomous physical/digital
objects augmented with sensing, pro-
cessing, and network capabilities. In
contrast to RFID tags, smart objects
carry chunks of application logic that
let them make sense of their local situ-
ation and interact with human users.
They sense, log, and interpret what’s
occurring within themselves and the
world, act on their own, intercommu-
nicate with each other, and exchange
information with people.
The combination of the Internet and emerging technologies such as near-
eld communications, real-time localization, and embedded sensors lets us
transform everyday objects into smart objects that can understand and react
to their environment. Such objects are building blocks for the Internet of
Things and enable novel computing applications. As a step toward design and
architectural principles for smart objects, the authors introduce a hierarchy of
architectures with increasing levels of real-world awareness and interactivity.
In particular, they describe activity-, policy-, and process-aware smart objects
and demonstrate how the respective architectural abstractions support
increasingly complex application.
Gerd Kortuem
and Fahim Kawsar
Lancaster University
Daniel Fitton
University of Central Lancashire
Vasughi Sundramoorthy
University of Salford
Smart Objects
as Building Blocks
for the Internet of Things
JANUARY/FEBRUARY 2010 31
Smart Objects as Building Blocks
The vision of an Internet of Things built from
smart objects raises several important research
questions in terms of system architecture,
design and development, and human involve-
ment. For example, what is the right balance for
the distribution of functionality between smart
objects and the supporting infrastructure? How
do we model and represent smart objects’ intel-
ligence? What are appropriate programming
models? And how can people make sense of and
interact with smart physical objects?
A key insight of our work is that the answers
to these questions are interrelated, so it doesn’t
make sense to attempt to answer each question
in isolation. Through practical experimentation
and by prototyping many generations of smart
objects, we identied three canonical smart-
object types (see Figure 1) that we believe rep-
resent fundamental design and architectural
principles: activity-aware objects, policy-aware
objects, and process-aware objects. These types
represent specic combinations of three design
dimensions that we’ll discuss later. Here, we aim
to highlight the interdependence between design
decisions and explore how smart objects can
cooperate to form an “Internet of smart objects.”
Smart Objects
for Industrial Workplaces
Our exploration of smart objects and the Inter-
net of Things is informed by the requirements
of industrial application scenarios — in partic-
ular, in the petrochemical and road construc-
tion industries. Our rst case study investigated
chemical storage at a processing plant, in partic-
ular, the use and handling of chemical drums;
2
the second case study looked at “road patching,”
a typical maintenance task aimed at repairing
defects in a road’s surface (see Figure 2a).
3
Although RFID technology is widely
deployed in many industries, its use in tempo-
rary and highly dynamic work environments
such as construction sites is severely restricted.
To overcome the handicap of an extensive exter-
nal infrastructure, we chose to convert existing
work objects such as containers and tools (pave-
ment breaker, drum roller, and wacker plate
compactor) into smart objects by augmenting
them with embedded sensor devices (based on
an ARM7 processor) and wireless capabilities
(following the 802.15.4 near-eld radio stan-
dard). The resulting smart work objects can
autonomously interpret sensor data and make
decisions, but also communicate and cooper-
ate with each other. To enable user input and
output, we equipped smart objects with a small,
embedded display and a set of buttons. In addi-
tion, we developed a wireless wearable device
that functions as a remote interface device for
smart objects (Figure 2b).
Smart-Object Typology
Through a multiyear collaboration with indus-
trial partners, we were able to build vari-
ous design alternatives for smart objects and
explore the smart-object design space in depth.
Although we deployed several hardware plat-
forms to accommodate increasing computa-
tional requirements and emerging standards,
we essentially kept the same hardware design
throughout. The key differences in our designs
can be found along the following three design
dimensions:
• Awareness is a smart object’s ability to
understand (that is, sense, interpret, and
react to) events and human activities occur-
ring in the physical world.
• Representation refers to a smart object’s
application and programming model — in
particular, programming abstractions.
• Interaction denotes the object’s ability to
converse with the user in terms of input,
output, control, and feedback.
Through iterative exploration and testing of
various designs, we discovered that the most
useful designs weren’t evenly spread through-
Workows
Representation
Rules
Functions
Interactivity
Policy-aware
Activity-aware
Process-aware
Awareness
Figure 1. Smart-object dimensions. We can see the three canonical
object types, activity-aware, policy-aware, and process-aware.
Internet of Things Track
32 www.computer.org/internet/ IEEE INTERNET COMPUTING
out the design space but clustered around the
three main object types we introduced previ-
ously (see Figure 1). Table 1 summarizes these
object types and how they relate to the three
design dimensions just introduced.
Activity-Aware Smart Objects
An activity-aware object can record informa-
tion about work activities and its own use. In
particular, we can characterize it as follows:
• Awareness. An activity-aware object under-
stands the world in terms of event and activ-
ity streams, where each event or activity is
directly related to the use and handling of the
object (pick up, turn on, operate, and so on).
• Representation. Its application model con-
sists of aggregation functions for accumu-
lating activities over time.
• Interaction. Activity-aware objects primar-
ily log data and don’t provide interactive
capabilities.
Activity-aware objects are the simplest of
the three types, and they already support inter-
esting smart-object applications. For the con-
struction case study, for example, we developed
a pay-per-use tool that uses sensors to record
data about the timing and duration of its use
and how workers handle it.
4
The tool converts
this usage data into a nancial cost gure,
which equipment rental companies can use to
realize a pay-per-use business model. The tool
also detects worker misuse (for example, drop-
ping the tool to the ground or overheating it)
and automatically takes into account necessary
maintenance and repair costs. (Most equipment
in the construction industry is rented on a con-
tractual basis, but rent prices depend only on
contract length.) Pay-per-use tools benet con-
struction companies as well because they sup-
port real-time cost capturing in the eld.
Technically, an activity-aware smart object
analyzes the data stream from its sensors,
uses recognition algorithms to detect activi-
(a) (b)
Figure 2. Road-patching case study. This study used (a) a smart object deployed at a road construction site. Workers
used (b) wearable user interface devices that showed personal health records containing information about a worker’s
exposure to hazardous equipment vibration.
Table 1. Summary of smart-object types.
Awareness Representation Interaction Augmentation Example
application
Activity-
aware object
Activities and usage Aggregation
function
None Time, state (on/
off), vibration
Pay-per-use
Policy-aware
object
Domain-specic policies Rules Accumulated
historical data,
threshold warnings
Time, vibration,
state, proximity
Health and safety
Process-
aware object
Work processes (that is,
sequence and timing of
activities and events)
Context-driven
workow model
Context-aware task
guidance and alerts
Time, location,
proximity,
vibration, state
Active work
guidance
JANUARY/FEBRUARY 2010 33
Smart Objects as Building Blocks
ties and events, and applies application-specic
aggregation functions. Further discussion of
usage-based pricing policies for smart products
appears elsewhere.
5
Policy-Aware Smart Objects
A policy-aware object is an activity-aware
object that can interpret events and activi-
ties with respect to predened organizational
policies. We can describe it within our design
parameters as follows:
• Awareness. A policy-aware object under-
stands to what extent real-world activi-
ties and events comply with organizational
policies.
• Representation. Its application model con-
sists of a set of rules that operate on event
and activity streams to create actions.
• Interaction. A policy-aware object provides
context-sensitive information about object
handling and work activity performance. In
particular, it can issue warnings and alerts
if workers violate policies.
We’ve used policy-aware object design to
develop health and safety-aware smart objects
for chemical storage and road construction sce-
narios. In the rst case, we developed a smart
barrel with embedded storage rules for various
chemicals.
2
Depending on temperature, vibra-
tions, and barrels’ relative proximity, it informs
workers about safety violations and prompts
them to take appropriate action. In our con-
struction case study, we developed a family of
vibration-aware tools that can monitor workers’
exposure to dangerous vibrations.
3
These smart
tools aim to minimize the occurrence of vibra-
tion white nger (VWF), a painful and poten-
tially debilitating disease caused by long-term
accumulative exposure to vibrations. The smart
tools carry an explicit model of legal health and
safety regulations, which state maximum daily
and average exposure levels.
6
The tools record
equipment use and send information to a work-
er’s wearable tag, where it’s stored as a personal
health log. The tag visually indicates current
exposure levels (Figure 3b) and, if vibrations
exceed legal limits, alerts workers.
Technically, a policy-aware object is an
activity-aware object with an added embedded
policy model. The user interface is an important
aspect of policy-aware objects; they not only
1. Checkout from Depot
1. Load on to Van
3. Transport
6. Load on to Van
[Conrm checkout]
[Proximity with a Van AND
Proximty Lost With Depot]
[Proximity With a Van]
[Proximity Lost With a Van]
5. Use6. Unload at Depot
7. Checkin to Depot
[Proximity With Depot]
[Conrm Checkin]
[Use]
User interaction required
Context condition
(a)
(b)
Figure 3. Smart objects in the eld. We designed and eld tested
(a) a pneumatic pavement breaker prototype that gathers data
about usage patterns and provides context-aware guidance during
construction work. The top left image shows the provisional
attachment of the sensor board to the pavement breaker; the
lower left image shows the sensor board. To model the tool’s
organizational process, we use (b) a workow that denes the
work activities in which the smart object is involved.
Internet of Things Track
34 www.computer.org/internet/ IEEE INTERNET COMPUTING
record and interpret sensed data, but they also
give users timely information. In this sense,
policy-aware objects are interactive systems.
Process-Aware Smart Objects
Processes play a fundamental role in industrial
work management and operation. A process is a
collection of related activities or tasks that are
ordered according to their position in time and
space. A process-aware object represents the
most accomplished of our three objects types;
we characterize it as follows:
• Awareness. A process-aware object under-
Related Work in Smart Objects
R
esearch on smart objects and the Internet of Things has
been going on for more than a decade and reaches back to
Mark Weiser’s original vision of ubiquitous computing. Bruce
Sterling recently popularized the idea of smart objects and the
Internet of Things; Sterling coined the term spime
1
to describe
a new category of space-time objects that are aware of their
surroundings and can memorize real-world events. Julian
Bleeker advocated a similar notion of blogjects (objects that
blog) in his “Manifesto for Networked Objects.”
2
This more
visionary work has been met by a growing body of technol-
ogy- and business-focused research on RFID, smart objects,
and smart products.
3
Roy Want and his colleagues augmented physical objects
with passive RFID tags so that they were uniquely identi-
able and information related to them could be presented to
their users.
4
Michael Beigl and his colleagues dened a smart
object as “an everyday artifact augmented with computing and
communication, enabling it to establish and exchange informa-
tion about itself with other artifacts and/or computer appli-
cations.”
5
Friedemann Mattern formulated in a similar way:
“Smart objects might be able to not only to communicate with
people and other smart objects, but also to discover where
they are, which other objects are in the vicinity, and what has
happened to them in the past.”
6
Norbert Streitz and his col-
leagues looked at smart objects from two perspectives: one
model has system-oriented, importunate smartness in which
smart objects can take certain self-directed actions based on
previously collected information; the other is people-oriented,
empowering smartness where smart objects empower users
to make decisions and take mature and responsible actions.
7
Most recent work on smart objects has focused on tech-
nical aspects (hardware platforms, software infrastructure,
and so on
8
) and application scenarios. Application areas range
from supply-chain management and enterprise applications
9
to
(home and hospital) healthcare
9
and industrial workplace sup-
port.
10–13
Human-interface aspects of smart-object technology
are just beginning to receive attention.
14
Yet design principles
and methods for smart objects that go beyond mere hardware
have yet to be explored. Our work on exploring the smart
object design space and identifying canonical smart object
types is a rst step in this direction (see also Fahim Kawsar’s
dissertation.
15
). In particular we view as paramount to holisti-
cally investigate sensing, modeling, and user interface issues.
References
1. B. Sterling, Shaping Things, MIT Press, 2005.
2. J. Bleecker, “A Manifesto for Networked Objects — Cohabiting with
Pigeons, Arphids and Aibos in the Internet of Things,” blog, 2006; http://
research.techkwondo.com/blog/julian/185.
3. F. Thiesse and M. Kohler, “An Analysis of Usage-Based Pricing Policies for
Smart Products,” Electric Markets, vol. 18, no. 3, 2008, pp. 232–241.
4. R. Want et al., “Bridging Physical and Virtual Worlds with Electronic Tags,”
Proc. ACM SIGCHI 99, ACM Press, 1999, pp. 370–377.
5. M. Beigl, H.-W. Gellersen, and A. Schmidt, “MediaCups: Experience with
Design and Use of Computer-Augmented Everyday Objects,” Computer Net-
works, vol. 35, no. 4, 2001, pp. 401–409
6. F. Mattern, “From Smart Devices to Smart Everyday Objects,” Proc. Smart
Objects Conf. (SOC 03), Springer, 2003, pp. 15–16.
7. N. Streitz et al., “Designing Smart Artifacts for Smart Environments,” Com-
puter, vol. 38, no. 3, 2005, pp. 41–49.
8. F. Siegemund, “A Context-Aware Communication Platform for Smart
Objects,” Proc. 2nd Int’l Conf. Pervasive Computing (PERVASIVE 04), Springer,
2004, pp. 69–86.
9. S. Konomi and G. Roussos, “Ubiquitous Computing in the Real World: Les-
sons Learnt from Large-Scale RFID Deployments,” Personal and Ubiquitous
Computing, vol. 11, no. 7, 2007, pp. 507–521.
10. M. Strohbach et al., “Cooperative Artefacts: Assessing Real World Situa-
tions with Embedded Technology,” Proc. Int’l Conf. Ubiquitous Computing
(Ubicomp 04), Springer, 2004, pp. 250–267.
11. G. Kortuem et al., “Sensor Networks or Smart Artifacts? An Exploration
of Organizational Issues of an Industrial Health and Safety Monitoring Sys-
tem,” Proc. Int’l Conf. Ubiquitous Computing (Ubicomp 07), Springer, 2007, pp.
465–482.
12. D. Fitton et al., “Exploring the Design of Pay-Per-Use Objects in the Con-
struction Domain,” Proc. 3rd European Conf. Smart Sensing and Context,
Springer, 2008, pp. 192–205.
13. C. Efstratiou et al., “Experiences of Designing and Deploying Intelligent Sen-
sor Nodes to Monitor Hand-Arm Vibrations in the Field,” Proc. 5th Int’l Conf.
Mobile Systems, Applications, and Services (MobiSys 07), ACM Press, 2007,
pp. 127–138.
14. L. Nelson and E.F. Churchill, “User Experience of Physical-Digital Object Sys-
tems: Implications for Representation and Infrastructure,” Smart Object Sys-
tems Workshop, in conjunction with Int’l Conf. Ubiquitous Computing (Ubicomp
05), 2005; http://elizabethchurchill.com/professional/pubs/Papers/2005
UbicompNelsonChurchill.pdf
15. F. Kawsar, A Document-Based Framework for User Centric Smart Object Sys-
tems, PhD dissertation, Dept. Computer Science, Waseda Univ., Feb. 2009.
