Ontology Support for Home Care Process Design

Jonas Bulegon Gassen, Alencar Machado, Lucin

eia Heloisa Thom and Jos

e Palazzo M. de Oliveira

Instituto de Inform

atica, Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, Brazil

Keywords:

Home Care, Ontologies, Business Process, Web Services, Process Design.

Abstract:

For the implementation of home care systems, using sensors and devices to provide assistance to the patient,

it is necessary to develop an easy way to conﬁgure each home, since differences between them exist and

a personalized one-by-one planning is expensive. To support this conﬁguration need we propose a Service

Oriented Architecture which uses process models to create service orchestrations to personalize each environ-

ment. Process design requires a good understanding of the application domain; business analysts interview

the domain experts and translate their understanding to process models. This modelling may be very complex

as the vocabulary used by the domain expert is very speciﬁc and difﬁcult to understand by process analysts.

Therefore, the process model elements (such as activities and role labels) can be named with inappropriate

terms. Moreover, the lack of domain understanding by business analysts increases the probability of errors

in the process design. In this paper we propose a methodology in which ontologies are used to support the

process design, expressing the modelling possibilities and also provides semantic to help the development

process. Altogether the approach aims to facilitate the design of each service orchestrations for home care

systems, by non specialists.

1 INTRODUCTION

The elderly population increased over the last years.

Thus, home care strategies show to be suitable mainly

because of cost and scalability reasons. Assuming

that intelligent homes will be offered, a problem of

scalability will appear since every patient and every

environment has their own particularities. Our re-

search challenge is to investigate how to facilitate the

design task by applying ontological support to deﬁne

the best suited workﬂows in a speciﬁc home project.

Our goal is to personalize homes from process

models, abstracting technical implementation issues.

This is equivalent to the Data Base development,

which starts from a conceptual schema that abstracts

any implementation issue. In our approach each pos-

sible action performed by the system is a Web Service

(WS). For example, to an actuator turn the lights on

there is an associated web service, to present some

information on the TV exists another WS, for a sen-

sor communicate the humidity of the ﬂoor and so on.

Based on an ontology, which expresses the possible

interactions (as turn the light on) and the involved par-

ticipants (as the home system or the caregiver), the

designer may model the situations for the patient con-

text.

As example: consider that a patient has Alzheimer

and sometime becomes agitated. In these cases the

system could interfere trying to calm the patient or

calling some related person. The ﬁrst step is to de-

tect the agitation behaviour (sensors are used in this

task, with WS), then the system checks the caregiver

location (WS). If the caregiver is at home, a message

is sent to him (WS), in other cases the system calls

an ambulance from the nearest care provider (WS).

In both cases the system checks the level of agita-

tion (WS) and if is high it suggests the use of drugs

(by message, WS). Otherwise it plays a music (WS)

that should calm down the patient. All tasks marked

with (WS) in the example occur through web services.

This model follows the process project concept and

may be implemented in a residential system.

2 BACKGROUND

Engineering is an expensive task in itself, but the con-

sequences of bad engineering are even worse, result-

ing, for instance, in the loss of the possibility to de-

ploy customized home care for a large public. In-

formation systems supporting the home care activi-

ties play a critical role in minimizing errors and de-

lays. Computer-supported tasks require the concur-

rent participation of multiple designers with a broad

Bulegon Gassen J., Machado A., Heloisa Thom L. and Palazzo M. de Oliveira J..

Ontology Support for Home Care Process Design.

DOI: 10.5220/0003995600840089

In Proceedings of the 14th International Conference on Enterprise Information Systems (ICEIS-2012), pages 84-89

ISBN: 978-989-8565-11-2

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

range of competences and a difﬁcult communication

barrier. The design task complexity is compounded

by its interdisciplinary character: the health care en-

gineering system has to assure design consistency be-

tween teams with different specialities that must work

as freely as possible. The approach, which is based

on a simple yet powerful conceptual model, combines

ontology and workﬂows to clarify the terms and con-

cepts and to facilitate the design process. Its fea-

tures are nowadays either supported by very expen-

sive or very specialized integrated packages, or sep-

arately performed by largely disintegrated database

managers, drafting systems and other software.

A home care design methodology must satisfy the

following requirements:

Abstraction. design on a given conceptual level

should be able to abstract design aspects concerning

more detailed levels;

Design Coordination. the conceptual model should

support the management of loosely coupled design

groups;

Conceptual Comfort. each concept of the model

should reﬂect established design habits naturally. In

the design of an instrumentation circuit familiar in-

strumentation symbols must be employed. When a

certain instrument is to be used at a certain location, a

set of instrument classes should be presented;

Generality. conceptual foundations must support a

broad scope of technologies, they must not be spe-

ciﬁc to any speciﬁc technology. Classes, types, rules,

e.g., must be deﬁned and customized by the user, not

“hard” pre-programmed into the design package;

Coupling between Disciplines. The design task is

multi-disciplinary: a home care environment is com-

posed of electrical, instrumentation, architectural,

leisure and other subsystems, which are designed by

different teams, but interact towards a common goal;

Knowledge Reuse. a home care environment should

support graphical drafting and design at large. For

drafting, a graphical interface is required. Design at

large includes engineering property veriﬁcation, us-

ing inference analysis.

A home care engineering project is an incremental

process starting from a coarse speciﬁcation and arriv-

ing at a detailed project. As projects are expensive the

reuse of former work is a necessity. The problem is

well suited for knowledge based and ontological sup-

ported solutions.

An ontology is a formal explicit speciﬁcation of a

shared conceptualization (Gruber, 1993). Conceptu-

alization refers to a model which describes concepts

(e.g. Patient, Caregiver, Room, etc.) and relation-

ships between these concepts of some domain (e.g.

Assists, hasCaregiver). The term “formal” means that

it has logics applied, which allows to avoid ambigui-

ties, also allows to apply inference mechanisms and to

check the model consistency. Furthermore, “shared”

means that this model is accepted by a group of peo-

ple. The ontologies are used to make the modelling

of business process easier, using its expressiveness to

restrict the model designer in some aspects. We use

the Web Ontology Language (OWL), which is a W3C

pattern.

A business process consists of a set of related ac-

tivities which are performed in a coordinated way.

These activities jointly execute a business goal. Each

business process is enacted by a single organization,

but it may interact with business processes performed

by other organizations (Weske, 2007). We understand

that situations of home care can be seen as a set of

activities that occurs in some sequence and in some

context, as well as business processes. For example,

if the patient gets agitated and a message is sent to

the caregiver it is a situation. If the patient gets agi-

tated and an ambulance is called, it is other situation.

Both occur in the same scenario, but in different con-

texts. The scenario presents all possibilities, the con-

text drives the execution performing a situation.

We use the Business Process Modelling Notation

(BPMN), which is a pattern developed under the co-

ordination of the Object Management Group (OMG,

2009). The business process will be used to create

service orchestrations, which controls the web ser-

vices execution. It allows personalized orchestrations,

based on the patient, environment and so on.

Web services can be deﬁned as a capacities

provider, composed by a body of logic which executes

them and by a service agreement, which expresses the

available capacities (Erl, 2007). Service orientation

comprises basically the use of independent services

as components to construct one composition with the

goal of accomplish some objective (Erl, 2007). In this

research, the home care environment includes a va-

riety of services such as: call an ambulance, send a

message to a caregiver, detect behaviour and so on.

3 RELATED WORK

Research on process design and ontologies have been

developed for more than ten years. One of the rea-

sons is that ontologies and structured vocabularies in

different domains help to make data understandable

by machines (Manzorr et al., 2007). However, most

of the existent approaches focus on building ontolo-

gies for the business process management domain as

well as in the use of ontologies to add more semantics

for the existent process model notations and execu-

OntologySupportforHomeCareProcessDesign

tion languages. Related to these points (Haller and

Oren, 2006), present an ontology that uniﬁes both in-

ternal and external business processes, based on vari-

ous existing reference models and languages from the

workﬂow and choreography domain. The authors ar-

gue that the interoperability problems in this domain

require an intermediate ontology to reduce the num-

ber of needed mappings.

An interesting approach regarding the semantics

of process design is proposed in the SUPER Project,

which has developed the Business Process Modelling

Ontology (BPMO) (Norton et al., 2009). The BPMO

enables the semantic annotation of high-level busi-

ness process models and it includes concepts to de-

scribe process behavior (workﬂow), activities and re-

lated organisational data. In the context of the SU-

PER project it was also proposed the Core Ontology

for Business pRocess Analysis (COBRA) which com-

prises a core terminology where business practitioners

can map domain-speciﬁc knowledge in order to anal-

yse their business processes (Pedrinaci et al., 2008).

For the intelligent homes there are many ongo-

ing researches as the Gator Tech Smart House (Helal

et al., 2005) which presents a service-oriented ap-

proach that allows to deploy new sensors, actuators

and devices in the home easily. They use the OSGi

(Open Services Gateway Initiative framework) to pro-

vide means for deploy new Java services. In our

approach the difference is that we use the service-

oriented approach based on SOAP, allowing the ser-

vice providers to choose how to implement the ser-

vices. Yet, we propose to use workﬂow managers to

execute the services and service compositions from

different providers, not only from the home.

Also an European project called Sm4all

(http://www.sm4all-project.eu/) which is in cer-

tain aspects very similar with our research. They

work in the domotics idea, using web services

(Catarci et al., 2011) to execute every available action

in the house, the actions (e.g. open the window

blinds) are provided by sensors, actuators and devices

in general. Furthermore, they work with the idea

of service compositions, the composition is made

automatically by the system that runs de residence,

from goals. They use the web service itself in the

composition, the abstraction that we use allows

using different services to execute the same action,

providing high decoupling and availability.

Considering these approaches, we consider that

the use of ontologies might facilitate the understand-

ing of the process domain, however, the process de-

sign and execution still constitutes a challenge.

4 ONTOLOGIES SUPPORT FOR

THE BUSINESS PROCESS

DESIGN

The proposed home care system consists of (i) care

providers (e.g. hospitals, general practitioners, phys-

iotherapy clinics, etc.), (ii) the residences of the el-

derly persons and (iii) any other entity that have in-

terest in providing services for those environments.

All participant entities communicate through web ser-

vices. So, the communication is possible directly be-

tween the participants or one of them (e.g. a home

care system provider) can act as intermediary, inte-

grating services provided by others. Examples of ser-

vices are:

• The residential system can provide a service for

patient localization;

• An intermediary system can provide a service for

choose the best care provider to call, based on in-

formation such as location, health plan, and so on;

• The hospital can have a service to inform the am-

bulances availability.

The available services require an orchestration to

accomplish the desired goals, process models will be

employed in this task. Having a knowledge model

(ontology) as base for creation of business process

models will make easier this task.

4.1 Methodology

To illustrate our proposal let’s consider the follow ex-

ample: (1) the patient goes to an appointment with

his physician that does the necessary prescriptions;

(2) The prescriptions are sent to the modelling de-

partment, where a non specialist person creates the

process model based on the prescriptions and a pre-

existing ontology; (3) The process(s) model(s) are

sent to the home system of the patient; (4) The res-

idential system adapts itself to the received model.

During the remain of the paper the second step will

be explained in more details.

Our methodology consists of two steps: to build

the ontology (by the home care system vendor) and

to design the business processes (by the process de-

signer) based on the former ontology. This paper does

not discuss the ontologies design, we assume that it

is already deﬁned and available in the web. The re-

sponsible for the ontology maintains the model well-

formed and consistent, and the updates are accessed

by the clients. The client’s users may be, for exam-

ple, a person who implements the prescriptions of

physicians to the automated residential context. In

ICEIS2012-14thInternationalConferenceonEnterpriseInformationSystems

this case, the person will probably have some domain

knowledge, but needs of an easy way to model it.

The ontology describes the concepts (eg. Resi-

dentialSystem, Caregiver, Patient, etc.) and the pos-

sible interactions between them in the environment

(e.g. DetectsAgitation, Alerts, Assist, etc). These in-

teractions are relationships between the concepts in

the ontology model, for example, the Caregiver (con-

cept) assists (interaction) the Patient (concept). This

way, the modeller selects the concepts and interac-

tions from the ontology, to create the activities, it is

not necessary to create them during the process de-

sign. Furthermore, the interactions that are compu-

tationally executed are linked to abstract representa-

tions of web services which allow to get information

about this action as the inputs, outputs and so on.

4.2 Activities

An activity is a unit of work that can be executed man-

ually or with the support of a business process man-

agement system. BPMN supports atomic activities

(tasks) or sub-process, that are ﬂows of activities.

According to (Mendling et al., 2010), little atten-

tion has been devoted to labelling the graphical con-

structs of a business process, in particular to activi-

ties, which are an essential part in this kind of mod-

elling. One of the main focus of our research is on

the atomic activities of business processes. Our con-

cerns are about the semantic of the labelling task of

activities. We propose a mark-up strategy instead of

free text labelling, which will make easier to under-

stand (computationally) the output generated from the

model. The activity will be composed of a triplet:

• Subject: performs an action (e.g. HomeSystem);

• Action: the action itself (e.g. notiﬁes);

• Object: suffers an action (e.g. Caregiver);

The examples create an activity for the residential

system (that manages the home) notiﬁes the caregiver

(person who takes care of the patient) about some-

thing that happened. The possible triplets are ex-

pressed in the ontology, the subjects and objects are

concepts, while the actions are relations between the

concepts. The process designer simply select the sub-

jects, actions and objects from the ontology to create

the triplets that are needed for the current modelling

task.

4.3 Actions

An action is something performed manually or com-

putationally during an atomic activity of a process.

Each action of the process model that is performed

for a computational entity will have an abstract rep-

resentation of a service or of a service capacity. For

the execution, the system can choose from the ser-

vice inventory the services that match with this rep-

resentation. This strategy separates the model of the

concrete (program codiﬁcation), providing dynamic-

ity for the home care system. If a service becames

unavailable, another one can be selected from the in-

ventory and so on. In order to describe the services, it

is possible to use the OWL-S ontology. Thus, describ-

ing the web services with OWL-S ontologies enables

the use of matching strategies as the one described in

(Maamar et al., 2011). Furthermore, other aspects can

be achieved by using these abstract representations of

services. For example, let’s say that the home sys-

tem needs to present some information (text) to the

patient. If the devices of the residence provide web

services, it is possible to match the abstract represen-

tation of the needed action against the services in the

home inventory.

An example refers to the similarity comparisons

of the abstract representation, verify if it is possible

to use services presenting different characteristics but

achieving similar goals. A problem occurs within the

TV, and the system needs to ﬁnd alternative ways to

present the information to the patient. The similar-

ity step does not found any other device that displays

information in the room, but found a service that re-

ceives text as input and reads it using a speaker, al-

lowing passing the message to the person. The com-

position remains the same because of the abstract rep-

resentation. During the model execution, the methods

can search the most suitable services, this provides a

high decoupling.

4.4 Ontology Advantages

In the present approach, ontologies support the cre-

ation of business process. The main advantages are

(1) To better understand the process model; (2) To

support business process reuse; (3) To facilitate dis-

cussion with other designers or specialists of the pro-

cess model domain; (4) To develop tools which can

make easier the process modelling task; (5) To rea-

son over the ontology and process to verify aspects as

most suitable activities for the patient; (6) To provide

characteristics of the actions (OWL-S), given more

control to the designer, over the activities; (7) To pro-

vide tracing capability as the activities have associ-

ated mark-ups. These advantages of ontologies add

more semantics for the process models. Also, they

help abstracting the technical issues in conﬁguring the

homes, since the designer selects the triplets from the

ontologies instead of program the services.

OntologySupportforHomeCareProcessDesign

Figure 1: BPMN Model of agitation behavior.

5 CASE STUDY

The scenario is presented by means of a process that

describes agitation behaviour. The process describes

an elderly person that presents mood and behaviour

changes, besides disorientation. Note that this pro-

cess is illustrative, was designed based on the litera-

ture and in group discussions of an Ontology Class

of a Public University located in Brazil (Siang Fook

et al., 2006). In real life this process can present dif-

ferent behaviour, the sample process present the situ-

ation discussed in late of Section 1. The whole sce-

nario of agitation in a BPMN model is presented in

Figure 1. In our research we will change the free text

describing an activity for a markup strategy, this make

easier the task of read the process computationally.

5.1 Modelling the Business Process

The modelling process is extended with our method-

ology; ﬁrst of all, there will be an ontology providing

the possibilities for the whole process modelling task.

The components used to form an activity are taken

from the ontology by selecting the desired concepts

or actions. To clarify the method Figure 2 presents

an illustrative image of an interface, in which the user

is developing the same model of Figure 1 using our

methodology.

As we can see in Figure 2, the ontology hierarchy

is loaded at the left, when the user creates a new ac-

tivity, he should choose a subject (concept) from the

tree, this concept will provide some possible actions,

which are presented on the top of the interface. After

chosen the action, only the concepts that make part

Figure 2: IIlustrative interface of a modelling tool.

of this action (objects) will be available on the hierar-

chy. These possibilities of representation of concepts

and interactions are expressed in the ontology and are

loaded by the modelling tool. In the ﬁgure it is possi-

ble to realize that a new activity was created (is high-

lighted), the subject chosen was “ResidentialSystem”

and the possible actions were loaded in the list on top

(the action “Send a notiﬁcation” is highlighted). The

texts that present the actions in the interface are labels

of the relations in the ontology.

The ontology provides the concepts and relations

for the domain in small granularity and the interface

provides the ability for aggregating those concepts

and relations to create activities (triplets). Also, the

modelling tool provides a manner to request new con-

cepts and relations to the ontology developers. The

compositions can be reused in various residences, not

all process must be personalized.

ICEIS2012-14thInternationalConferenceonEnterpriseInformationSystems

As the quantity of concepts and relations can grow

fast, depending on the domain, strategies based on

patterns for suggesting the next process creation step

will be included. In another paper is described a set

of workﬂow activity patterns which refers to descrip-

tions of recurrent business functions (Thom et al.,

2009). Some of these patterns can imply in the re-

striction of the next possible activities, for example,

after the use of the pattern X, only activities of the

pattern Y can be used. The modelling tool exports

an XML ﬁle that may be processed by the residential

control system.

6 CONCLUSIONS AND FUTURE

WORK

It is very important for the home care idea that the

systems developed are scalable in the dimension of

multiple locals since one of the main problems for

deploying home assistance is the increasing number

of elder people living alone. If the developed home

care systems will require the same number of people

as the present systems, the implementation of wide

spread home care will be compromised.

This paper described an approach that intends to

reduce the need of a huge number of new person-

nel for the design and implementation of home care

systems. The proposed methodology provides activ-

ities with more semantic to be employed in the pro-

cess model, making easier to understand this model

for persons as well being interpretable by automated

devices. This modelling approach can also be ap-

plied for other domains; we focus on home care in

this work. The use of a service-oriented system is

also interesting since services can be added, removed

or updated enabling the system to be context aware.

We are working to ﬁnish the prototype of the tool

for modelling the processes based on ontologies. Af-

ter that, we will work in a prototype for executing the

processes, probably based in an execution language

as the Business Process Execution Language (BPEL).

As future work, we intend to apply more deeply

the BPMN notation, in order to verify how much in-

heritance from ontologies may be applied to model

the process. Moreover the BPMN notation doesn’t

impose much focus in the process execution and a re-

lated study should be done.

ACKNOWLEDGEMENTS

This work has been partially supported by Conselho

selho Nacional de Desenvolvimento Cient

ıﬁco

e Tecnol

ogico (CNPq) and Coordenac¸

ao de

Aperfeic¸oamento de Pessoal de N

ıvel Superior

(CAPES), Brazil.

REFERENCES

Catarci, T., Di Ciccio, C., Forte, V., Iacomussi, E., Mecella,

M., Santucci, G., and Tino, G. (2011). Service com-

position and advanced user interfaces in the home of

tomorrow: The sm4all approach. Ambient Media and

Systems, 70:12–19.

Erl, T. (2007). SOA Principles of Service Design (The Pren-

tice Hall Service-Oriented Computing Series from

Thomas Erl). Prentice Hall PTR, Upper Saddle River,

NJ, USA.

Gruber, T. R. (1993). A translation approach to portable

ontology speciﬁcations. Knowledge Acquisition,

5(2):199 – 220.

Haller, A. and Oren, E. (2006). A process ontology to rep-

resent semantics of different process and choreogra-

phy meta-models. Digital Enterprise Research Insti-

tute DERI Galway, (DERI 2006-02-03):19.

Helal, S., Mann, W., El-Zabadani, H., King, J., Kaddoura,

Y., and Jansen, E. (2005). The Gator Tech Smart

House: a programmable pervasive space. Computer.

Maamar, Z., Santos, P., Wives, L., Badr, Y., Faci, N., and

Palazzo M. de Oliveira, J. (2011). Using social net-

works for web services discovery. Internet Comput-

ing, IEEE, 15(4):48 –54.

Manzorr, S., Ceusters, W., and Rudnicki, R. (2007). Im-

plementation of a referent tracking system. IJHISI,

2(4):41–58.

Mendling, J., Reijers, H., and Recker, J. (2010). Activity

labeling in process modeling: Empirical insights and

recommendations. Information Systems, 35(4).

Norton, B., Cabral, L., and Nitzsche, J. (2009). Ontology-

based translation of business process models. In Pro-

ceedings of the 2009 Fourth International Conference

on Internet and Web Applications and Services, Wash-

ington, DC, USA. IEEE Computer Society.

OMG (2009). Omg business process modeling notation

v1.2. Technical report, Object Management Group.

Pedrinaci, C., Domingue, J., and de Medeiros, A. K. A.

(2008). A core ontology for business process anal-

ysis. ESWC’08, pages 49–64, Berlin, Heidelberg.

Springer-Verlag.

Siang Fook, V. F., Tay, S. C., Jayachandran, M., Biswas,

J., and Zhang, D. (2006). An ontology-based context

model in monitoring and handling agitation behaviour

for persons with dementia. PERCOMW ’06, Wash-

ington, DC, USA. IEEE Computer Society.

Thom, L. H., Reichert, M., and Iochpe, C. (10 August

2009). Activity patterns in process-aware informa-

tion systems: basic concepts and empirical evidence.

International Journal of Business Process Integration

and Management, 4:93–110(18).

Weske, M. (2007). Business Process Management: Con-

cepts, Languages, Architectures. Springer-Verlag.

OntologySupportforHomeCareProcessDesign