![](bg7.png)
equipped with Tablets, PDAs, smart dispensers, vital
signs checkers and sensors. These sensors are able to
measure the environmental conditions (e.g. tempera-
ture and humidity).
Nancy is a professional caregiver and is responsi-
ble for creating and managing the homecare services.
Among the services configured by Nancy there is an
alert to remind John to take his medicine at the correct
time and to add a new medicine prescribed by Jonh’s
physician in a dispenser.
When a prescription is registered, it is associated
with an event, controlled by a timer, which is respon-
sible for initiating the process of notification. Then,
the system informs the patient in which dispenser the
medicine is stored. The system must adapt its ser-
vices pro-actively, at runtime, according to the pref-
erences or needs of users. For example, John prefers
to receive alerts on his mobile device without volume
when he is accompanied. In case of frequent ignored
reminders (e.g. due to the hearing impairment of the
John), the system can send a message to Nancy to re-
configure the services to get John’s attention.
When Nancy arrives at the house, the sensors de-
tect the presence of a relevant instance of a class.
Then, monitoring and data input module is notified
and inserts information about her into an XML docu-
ment and sends it to the OntoHC module.
After receiving the information, the OntoHC mod-
ule uses SQWRL queries to check if the instances de-
tected are already mapped in the ontology. In affir-
mative case, it checks what actions could be triggered
and informs the Notification module. However, if the
ontology does not have the classes, a new one must be
created with the necessary instances, and the current
should be stored in the repository, as seen in section
5.2. For this case study, we consider that the two in-
stances (Nancy and John) are mapped in the ontology.
When the event is triggered, the OntoHC mod-
ule searches for information about the device and dis-
pensers of the medicine. The system creates a XML
document with the necessary information for the mo-
bile application show a reminder to John. Then, the
file is forwarded to the notification module, respon-
sible for sending the information about the medicine
and the dispenser as a reminder to a device next to
John. After display the notification, the system checks
if the patient performed the task, within a range of
pre-set time, defined by Nancy. This information can
be captured by the sensors.
If the patient do not perform the task, the system
redirects the notification to the Nancy’s device and
she will be able to help him. Finally, the ontology of
the current context is updated, indicating that the task
was performed (patient took the medicine). When the
caregiver assists the patient, he can inform the com-
pletion of the task manually.
We can conclude that the system helps to reduce
problems of human failures (e.g. forgetting medica-
tion). The patient is able to keep his autonomy during
the treatment, reducing the intervention by the care-
giver.
7 CONCLUSION
Nowadays the hospitals are facing a problem of de-
mand of their services, where in some cases it is
higher than what they can support, creating queues
or a low satisfaction by the patients. One way to help
solving this problem is to take the services offered by
health care providers to the house of patients. Home-
care environments are characterized by being a do-
main with all the necessary infrastructure to the pa-
tient to be treated in his own house, without the need
of going to the hospital.
The development of pervasive applications helps
to ensure that the services provided to patients in their
houses have the same quality as those offered in hos-
pitals. Considering this, we first presented an archi-
tecture to be used as basis to develop pervasive sys-
tems for homecare environment and also an ontology
to represent the knowledge that this kind of domain.
In this paper we described the management of the
ontology, showing the workflow process of the mod-
ule responsible by the creation of ontologies and how
the inference on it is performed. The model developed
was validated through a case study. Systems devel-
oped from the architecture can help users in different
scenarios ranging from small tasks like remember pa-
tients of their daily activities or cognitive games for
brain exercises, until more complex applications such
as to assist the professionals during their duties, sug-
gesting actions to be taken by them.
We intend to continue this project describing the
other modules of the architecture through workflow
processes models and focus on some aspects like pri-
vacy and safety of information. Until this part of the
project we assume that the homecare environment has
the necessary infrastructure that allows the develop-
ment of pervasive applications considering these as-
pects. Then, we intend to build a real homecare envi-
ronment to develop tests and simulate real situations,
where we will be able to analyse aspects like perfor-
mance of the network (e.g. the time between the de-
tection of a new instance by sensor until the creation a
new ontology to be used by the system). With this, we
will be able to make a better analysis of the system,
and then make it available to real users.
WorkflowModelforManagementofOntologyofHomecarePervasiveSystems
247