CONTROL MODEL OF DOMOTIC SYSTEMS
BASED ON ONTOLOGIES
Pablo A. Valiente-Rocha and Adolfo Lozano-Tello
Universidad de Extremadura, Área de Lenguajes y Sistemas Informáticos
Escuela Politécnica, Campus Universitario s/n, 10071, Cáceres, Spain
Keywords: Ontologies, SWRL, Ambient Intelligence, Domotic, IntelliDomo.
Abstract: In the present paper we introduce the model process of an expert system for the control of domotic
installations based on ontologies and SWRL rules. From the domotic system database where the attribute
values of each device are stored, a background process converts these values into instances of the ontology
representing the system. From these instances, a software application -known as DomoRules- allows
creating production rules in SWRL language that will be useful to regulate the system. IntelliDomo draws
inferences from the ontology and the SWRL rules by using behaviour parameters previously indicated by
the user, so the state of the physic devices in the domotic system can be modified in real time.
1 INTRODUCTION
The use of Ambient Intelligence (AmI) is one of the
areas which are rapidly gaining importance in the
application of intelligent systems in companies and
homes. One of the characteristics AmI systems seek
is the gift of interaction with the user so that
communication is as far as possible natural and the
configuration of preferences takes less time for the
user. AmI systems use behaving policies able to
draw deductions with the information provided by
certain devices and the represented rules. Moreover,
they make decisions regarding the management of
the installation.
The use of ontologies as representing bases,
together with the representation of rules in SWRL in
AmI systems, will provide the most precise
definition of the taxonomy of physical devices that
may exist in a system, the attributes of these devices
and the possible relations among them. Furthermore,
these representations will be more reusable by other
users and will favor the completion of the
classification of domotic components, the useful
information to be represented upon these
components and the connection rules that will allow
deducting new information regarding the values of
other components.
The amount of literature about intelligent
domotic environments, which makes use of
ontologies as representing bases, is poor. DomoML
is one of the main reference works (Sommaruga et
al, 2005)(Furfari et al, 2004); it is a markup
language focused on defining a communication
method among domotic devices.
Related to the previous project, DogOnt (Bonino
& Corno, 2008) proposes a system able to recognize
the devices that comprise the domotic environment
automatically. DogOnt uses SWRL language to
define rules that allow completing the ontology
model which represents the domotic environment.
In these projects ontologies are used as
representing bases of domotic devices and SWRL
rules are used to maintain consistency in the
information of the system components values.
However, these production rules are not used to
control general autonomous and real time
functioning of the system.
The present paper introduces IntelliDomo, an
AmI system based on ontologies for the control of
domotic systems. It uses domotic components and
its state values represented as instances of an
ontology, and takes advantage of the power of the
production SWRL rules specified by the user in
order to change the state of the system components
in real time. In section 2 of the present paper, we
will describe the architecture of IntelliDomo and, in
470
A. Valiente-Rocha P. and Lozano-Tello A. (2010).
CONTROL MODEL OF DOMOTIC SYSTEMS BASED ON ONTOLOGIES.
In Proceedings of the 2nd International Conference on Agents and Artificial Intelligence - Artificial Intelligence, pages 470-473
Copyright
c
SciTePress
Figure 1: Overview of IntelliDomo Arquitecture.
its subsections, we will describe each of the
components and tools of the model software.
Finally, section 3 is dedicated to conclusions and
future lines.
2 MODEL PROCESS FOR
THE CONTROL OF DOMOTIC
INSTALLATIONS
BASED ON ONTOLOGIES
IntelliDomo is an expert system able to control the
components of a domotic system automatically and
in real time. It is based on an ontology called
OntoDomo which contains the information about the
devices of the system, on production rules in SWRL
language which shape its behaviour and on a series
of interconnection software tools, rules design and
inference motor (see Figure 1).
The main characteristic of IntelliDomo is the
ability of making decisions and reacting to the
changes that arise in the elements of the domotic
system. These decisions are taken based on the state
of the system elements represented in OntoDomo
and also on a set of established rules, expressed in
the SWRL rules definition language. These rules can
be modified by the user with the use of the
DomoRules tool presented in the following
subsection. In order to draw knowledge deductions,
IntelliDomo uses Jess inference motor, this tool will
allow inferring the concepts of the ontology and
obtaining new data according to what it is stated in
SWRL rules.
In order to manage this knowledge, IntelliDomo
is built upon an ontology which concepts are related
to the domotic components. The ontology has been
modeled to be in sync with the physic devices that
form a domotic environment, so that it can store its
outstanding values and properties.
IntelliDomo has been designed to interact with
an existing domotic database (DomoBD), where the
state and values of the domotic components are
updated in real time. This database can be obtained
throughout the QDSConnect software module that
has been developed by the research team from the
Universidad de Extremadura, Quercus. This daemon
provides the physic connection with EIB/KNX
bus(http://www.knx.org). QDSConnect uses
FALCON library to translate values of the physic
devices into the relational database system and to
detect any changes that may occur in the mentioned
database in order to modify it straight in the physic
devices.
Therefore, IntelliDomo can make deductions and
update the domotic system instantly by transferring
the new values from its hardware components to the
database. DomoBD is the link between IntelliDomo
and the physic domotic system. The information of
this database is constantly translated into instances
of OntoDomo by DBConnection Module. Similarly,
when Intellidomo fires the correspondent rules,
DBConnection Module updates DomoBD database,
which will entail the respective changes of state in
the physic devices.
2.1 OntoDomo Ontology
Nowadays, ontologies are the most used way of
knowledge representation in different business or
research projects in fields such as databases,
intelligent information integration, cooperative
information systems, information retrieval,
electronic commerce, enterprise application
integration, and knowledge management (Hepp,
2008). IntelliDomo uses ontologies as a representing
base of domotic elements that constitute the system.
These elements are stored in OntoDomo.owl, which
CONTROL MODEL OF DOMOTIC SYSTEMS BASED ON ONTOLOGIES
471
is responsible for reusing the knowledge of the rest
of ontologies constituting IntelliDomo.
IntelliDomo allows the users to establish
configurations able to guide the system behaviour
and that are adapted to its needs and preferences. In
order to do so, it uses an ontology that allows storing
the configuration into profiles and to make
inferences and reasoning able to incorporate new
knowledge to the system. The ontology
Preferences.owl is used with this end. Here, each user
will be able to store his/her personal information for a
certain situation: security, temperature, comfort... A
series of variables defining the user's preferences can
be established for every situation. For instance, as
for temperature, the user will be able to specify that
‘hot’ ranges between 25 ºC and 36 ºC, consequently,
these values will be used in the corresponding
production rules.
Besides, the ontology Preferences.owl controls
the so-called ‘system variables’. These are control
variables which could be used in SWRL rules to
work with them and draw inferences. These
variables work as global variables that will be shared
by each user and known by the inference system in
order to use them as a reference for certain
decisions. These variables are identified with the
symbol ‘#’ prior to the name.
2.2 DomoRules Tool
SWRL is based on OWL ontology language, more
specifically on the OWL-DL branch. It is a language
able to build up rules to perform reasoning about the
instances of an OWL ontology (OWL Individuals)
and infer new knowledge about them.
DomoRules is a tool developed in Java that
facilitates the creation of SWRL rules throughout
API SWRLFactory. It provides a graphic interface
appropriate to assist in the construction of these
rules. It allows obtaining from OntoDomo all the
necessary elements to build a SWRL rule (classes,
instances, properties), to use SWRL variables and
system variables that refer to the values defined by
the user and finally to add SWRL language built-ins
to the rules.
With DomoRules we seek to build SWRL rules
easily without knowing the syntax of this language.
All the elements constituting OntoDomo.owl
ontology are introduced to the user in the interface,
so that, by simply choosing a class or an instance, all
the properties it has will be listed and the user will
be able to pick the one he/she needs and establish its
values in order to build the rules.
DomoRules is born aiming to offer the user a
simple SWRL rules creating application. The task of
designing an intuitive interface to abstract the
SWRL language user is not easy because SWRL is a
powerful language with a very rich vocabulary that
provides high interaction with all its elements.
Accordingly, by using DomoRules the user will be
able to produce rules that shape the system
behaviour that could be used by IntelliDomo.
Figure 2: DomoRules creation rules interface.
2.3 IntelliDomo Module
IntelliDomo is the main control module of the
system using the represented elements and the
different modules which constitute the system. The
information about configuration and location of the
system can be observed graphically on a plan and,
when the state of a device changes, a message alerts
appearing upon the device icon. At the same time
these events happen, it stores a set of log files where
each of the actions carried out by the system and all
the inferences it makes are detailed. It uses the
DBConnection module to get connected to the
different databases that constitute the application,
which can be either DomoBD or the private
databases IntelliDomo uses to store its own data.
3 CONCLUSIONS AND FUTURE
LINES
Ontologies provide an appropriate kind of
representation for identifying types and
characteristics of domotic devices. Furthermore,
production rules expressed in SWRL language allow
establishing relationships among these domotic
devices to shape the integral behaviour of a domotic
ICAART 2010 - 2nd International Conference on Agents and Artificial Intelligence
472
installation.
In the present paper, we introduce IntelliDomo.
On the one hand, it is based on an ontology -known
as OntoDomo- where all types of domotic devices,
together with their useful characteristics, are
represented; and, on the other hand, it is based on
production rules in SWRL language defined
according to the users’ preferences and needs. With
these knowledge bases, IntelliDomo allows
managing the control of the domotic system itself.
The state of the elements that comprise a determined
domotic installation is continuously read from the
database where its values are stored and translated
into instances of OntoDomo ontology. With this
information, together with the rules defined by the
user, IntelliDomo’s inference engine would fire the
appropriate rules that will change the state of the
system devices.
In addition to the IntelliDomo control module,
we have developed DomoRules application, which
works as a SWRL production rules generating
wizard. The interface design has been developed
aiming to be simple and intuitive for every user. As
a result, the user may only select the type of
component (a concrete physic component in the
system) and the appropriate values to build the
condition of the rule antecedent (body).
Nevertheless, we are inclined to provide
IntelliDomo with a learning module where
production rules would be modified according to the
historical procedure of the users. The intention is
that, meanwhile the user interacts with the system or
some determined values are obtained under certain
conditions, the rules can change or their execution
precedes other rules. Therefore, once the user has
established the reasonable importance of behaviour
aspects, rules can be modified in order to adapt the
system behaviour according to the importance of
these criteria.
ACKNOWLEDGEMENTS
This work has been developed under support of
Junta de Extremadura Project (PDT08A023)
REFERENCES
Horrocks I., Patel-Schneider P., Boley H., Tabet S.,Grosof
B., and Dean M. SWRL: A semantic web rule
language combining OWL and RuleML.
http:// www.w3.org/Submission/2004/SUBM-SWRL-
20040521/, Mayo 2004.
Sommaruga, L.; Perri, A.; Furfari, F.; DomoML-env: a
ontology for Human Home Interaction. Proceedings
of SWAP (2005)
Furfari, F.; Sommaruga, L.; Soria, C.; Fresco R.
DomoML: the definition of a standard markup for
interoperability of human home interactions.
Proceedings of the 2nd European Union symposium
on Ambient Intelligence (2004)
Bonino, D.; Corno, F.; Dogont-ontology modeling for
intelligent domotic environments. 7th International
Semantic Web Conference (2008) pp. 790-80
Hepp, M.; De Leenheer, P.; de Moor, A.; Sure, Y. (Eds.)
Ontology Management. Semantic Web, Semantic
Web Services, and Business Applications,
Vol. 7.Springer 2008
CONTROL MODEL OF DOMOTIC SYSTEMS BASED ON ONTOLOGIES
473