446 ◾ Slobodanka Tomic et al.
Services within the SESAME architecture provide dierent functionalities
at dierent interfaces. For example, smart meter data is published by the AMI
provider through an external SOAP-based Web service. On the same side of
SESAME, the Web service client invokes this service and updates the knowl-
edge base. Sensors, appliances, and displays are also implemented as service-based
information publishers and consumers. As a home automation system is inher-
ently an event-based system, each service interface also implements a notication
passing capability for service-based interactions of users and energy providers or
grid operators.
e SESAME service architecture is based on the OSGI. To account for
dynamic changes in the environment and user expectations, SESAME designs
exible life-cycle management of ontologies and services. is provides means for
dealing with dynamic additions of new system components (new types of devices)
or adding a new user policy or workow, a new energy provider, or a new home.
e system allows “plug-in” ontologies and ontology evolution and selection driven
by user needs.
16.4.3 Ontology Design
SESAME uses an ontology-based modeling approach to describe an energy-aware
home and the relationships of the objects and actors within its control scenario. e
SESAME ontology provides a hierarchy of concepts to model the automation and
energy domains and is specied in OWL. e main components are the automa-
tion ontology, meter data ontology, and pricing ontology.
SESAME Automation Ontology (Figure16.3) includes a number of general
concepts such as Resident and Location, and concepts in the automation and in
the energy domain, such as Device, and Conguration. e Device class has sub-
classes modeling Appliances, Sensors, or simple message-based User Interface (UI)
devices. New energy-related properties in the Device model are consumption per
hour, peak power, and the switch on/o status as well as the required state “to be
switched on/o.” For an Appliance, we also introduce the property “canBeStarted”
which models the state of the devices for which activation can be scheduled. For
example, after a user lls in his washing machine he congures it (via a new UI)
with the “canBeStarted” set to “true,” and with the time interval within which the
washing task should be accomplished.
To model dierent types of control functionality, the SESAME ontology intro-
duces the Conguration class, which has two subclasses: Activity (or automation
activity) and EnergyPolicy. An Activity connects Appliance, Sensor, and UI Device
into a joint task. A ContextBased Activity can provide regulation of dierent types,
e.g., regulation on time, occupancy of location, or threshold value. For this pur-
pose it includes properties including thresholds and scheduled times. An example
of a ContextBased Activity would be HeatingRegulationBedroom which would
connect TemperatureSensor in the Bedroom and Heater. is Activity would be