Pervasive Ambient Intelligence Platforms in the IOT Era based

on a Ubiquitous User Model Ontology

An Implementation Account

Alfio Costanzo, Alberto Faro and Daniela Giordano

Department of Electrical, Electronics and Computer Engineering

University of Catania, viale A.Doria 6, 95125, Catania, Italy

Keywords: Pervasive Information Systems, Ambient Intelligence, Ubiquitous User Model Ontology, Smart Cities.

Abstract: This paper presents how ambient data integration is obtained in Wi-City, i.e. a project promoted by the

Regional Government of Sicily that aims at supporting mobile people activities by means of intelligent

applications able to generate personalized recommendations that take into account both personal and context

parameters. The paper shows how this is made possible by the decreasing cost of the monitoring systems in

the IOT era and by the availability of ontology engineering methods to data integration. In particular, aim

of the paper is to illustrate a pervasive platform consisting of environmental sensors readily installable on

the city and body sensors easily wearable by people that cooperate by means of an ambient data ontology to

better support the user decisions in a smart city. Examples of how embedded consumer electronics products

are used to monitor the user ambient and how developing an effective ambient data ontology are illustrated

to give an implementation account of the proposed platform.

1 INTRODUCTION

In a smart city, the citizens are able to use the

technologies that improve the relevant aspects of

their life (Aoun, 2013). A step beyond the smart city

is the intelligent city scenario where an open and

interoperable city information platform is able to

support ubiquitous decision making by using

intelligent systems (Berthon, 2011). In this scenario

ambient information plays an important role. Indeed,

at the basis of an effective decision support system

there is the information coming from devices able to

recognize the people activities, to measure the

parameters of the systems in which the people

carries out its activities, including traffic and

environment, and to monitor the health conditions.

Today, the various city life aspects are mainly

managed by separate decision support systems and

consequently the ambient intelligence is partitioned

into sub-ambients that don't communicate between

them, e.g., car navigators help drivers only to find

the best path to the desired destination, whereas

specific e-health applications send alarms to rescue

the users involved in accidents.

However, the increasing diffusion of mobile

devices makes possible, in principle, to inform the

users about the most suitable actions using the same

terminal device independently on if they are at

home, at office or are walking or driving.

Also, the more and more decreasing cost of the

monitoring systems in the era of IOT (Internet Of

Things) makes possible to support the user activities

by means of intelligent applications able to generate

recommendations that take into account both

personal and context parameters. Indeed, these

ambient data may be monitored by means of

embedded consumer electronics products provided

with network identifiers and cooperating between

them under the supervision of an Ambient

Intelligence (AmI) software resident on the user

mobile and/or in some remote city server.

For example fig.1 shows the pervasive e-health

service envisaged in (Acampora, 2013) for an

ubiquitous rescue of people. It is based on the

interconnection of different AmI technologies

involving body sensors, emergency response and

hospital equipments.

Although such picture has the merit of

envisaging the interconnection of the data collected

by different AmI technologies to give rise to a

ubiquitous system at urban scale, it underestimates

the critical problems to be solved for the real

Costanzo A., Faro A. and Giordano D..

Pervasive Ambient Intelligence Platforms in the IOT Era based on a Ubiquitous User Model Ontology - An Implementation Account.

DOI: 10.5220/0004812900720078

In Proceedings of the 4th International Conference on Pervasive and Embedded Computing and Communication Systems (PECCS-2014), pages 72-78

ISBN: 978-989-758-000-0

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

Figure 1: A general perspective of a pervasive e-Health

system hiding critical implementation issues.

interoperability of the data collected by the different

consumer electronics devices.

An ubiquitous architecture able to manage all the

mentioned information sub-systems has been

proposed by the authors in previous papers where all

the mentioned data are integrated by an ontology

based approach and a fuzzy logic based engine is

adopted to support the decisions of the users and of

the rescuers, e.g., (Costanzo, 2013a).

The original aim of Wi-City was to support

mobility and logistics activities of citizens. Then, the

monitoring system was mainly dedicated to collect

traffic data by in situ technologies, e.g., (Leduc,

2008), (Faro, 2008), (Faro, 2011a). Such data are

sent to the main city server where they are stored in

XML/RDF format (http://www.w3.org/RDF). Also,

personal data giving a general description of the

current user status and task are stored on the user

mobile so that the Decision Support System (DSS)

implemented on the server or in the user's mobile

may be able to suggest the best path to destination or

the best delivery cycle taking into account the traffic

conditions and the general user status. Data of city

interest stored on public or private databases may be

used by the DSS to identify the service most suitable

for the mobile-device-using users.

Aim of the paper is to improve this architecture

by taking into account the local and global ambient

data collected by an ubiquitous monitoring platform

consisting of cooperating sensors readily installable

on the city and easily wearable by people to better

support the user decisions.

In particular, in sect.2 we show how the

functional Wi-City architecture has been extended to

include the ambient monitoring functions. In

addition, the consumer electronics products and

their interconnection to implement the above Wi-

City extension are briefly illustrated. Sect.3

discusses the data ontology for ambient data

integration at city level taking into account the

General User Model Ontology (Heckmann et al.,

2005) and its extension (Costanzo, 2013b).

2 AMBIENT DATA IN WI-CITY

Fig.2 shows how the Wi-City architecture has been

extended to take into account the ambient data. In

particular, suitable applications have been developed

to measure the indoor environment parameters and

the physiological variables so that the DSS resident

on the mobile provided with Flash Builder based

software version (Corlan, 2009) may take into

account the user's health status and regulate the

indoor environment, e.g., (Costanzo, 2014). The

main weather parameters are collected by a

distributed monitoring system supervised by the Wi-

City server to inform the user about critical weather

conditions that influence the outdoor activities.

Figure 2: Wi-City architecture including ambient data

Fig.3 shows the implementation structure adopted in

Wi-City to monitor the main parameters of the

outdoor environment. The interested reader may find

the ones dealing with e-health and indoor monitoring

in (Costanzo, 2014).

The internal structure of this station allows us to

point out that, apart the mechanical devices devoted

to measure the rain quantity and wind

intensity/direction that have dimensions of about ten

centimetres (see fig.4a), the other devices measuring

pressure, temperature, humidity and gas emission,

may be embedded into a system of very limited

dimensions, as shown in fig. 4b.

People at

Smart Home

Mobile Body

Sensors

Smart Hospital

Smart

Transportation

Medical Personnel

Emergency

Services

TRAFFIC

MONITORING

WEATHER

MONITORING

WI-CITY

SERVER

PervasiveAmbientIntelligencePlatformsintheIOTErabasedonaUbiquitousUserModelOntology-AnImplementation

Account

Figure 3: Wi-City environment monitoring station.

Figure 4: Environment monitoring station: apparatus to

measure rain quantity and wind direction/intensity on the

left, and Arduino controller on the right.

To transform the sensors of such station into things

of internet, they have been supervised by an Arduino

board (Banzi, 2009). Fig.5b shows how this board

collects all the data coming from the environment

sensors and sends them, through a GPRS/GSM

shield, to the main Wi-City server where they are

available to the users. This station has been provided

with a Arducam to make available on internet a

picture of the zone involved in bad weather or

congested traffic conditions.

The users' mobiles may access directly the data

stored on the stations so that the DSS resident on the

mobiles may obtain very fast the ambient data

without overloading the main server. In other words,

the modularity of the system allows the user's

mobiles to work cooperatively with the monitoring

stations without the server intervention, thus

increasing the time performance of the system.

The chosen implementation architecture is fault

tolerant since the system functions show a graceful

degradation in case of a failure of a component.

Also, it guarantees high reliability and accuracy at

low cost since cheaper electronics components able

to work with a satisfactory continuity and precision

are widely available on the market.

On the contrary, for what concerns the

interoperability, although the data collected by the

Arduino controllers are sent in JSON format

(http://www.json.org) that can be easily converted

into XML/RDF statements, this does not guarantee

per se the data interoperability.

Indeed, to this aim, it is required that not only

every measured data is stored by a triple (subject,

predicate, object), e.g.: (device_id_01, temperature,

25°), but also that the terminology used, i.e., the

predicate temperature and related properties, should

be known by both the Arduino controllers and the

DSSs so that they may work cooperatively. A

suitable ambient data ontology is sketched in sect.3.

3 AMBIENT DATA ONTOLOGY

Currently, there is no data ontology proposed by the

standardization organization, e.g. the World Wide

Web Consortium (W3C), to represent the data

featuring all the aspects relevant for the citizen

activities, e.g., health, environment and preferences.

However, this should not prevent the use of the

ontological approach to data integration. Indeed, one

may develop a suitable ontology to manage the

relevant aspects that a DSS should consider to issue

its recommendations to support mobile people in a

smart city and publish this ontology using a suitable

editor, e.g., Protégé (http://protege. stanford.edu) so

that it may be used by any remote processes to

extract the XML/RDF data stored in the mobiles or

in the main server using JSON queries.

Also, this solution allows the remote DSSs to

issue SPARQL queries (http://www.w3.org/TR/ rdf-

sparql-query) if the data are formatted by means of

the Ontology Web Language (OWL) defined in

(http://www.w3.org/2004/OWL).

A suitable ontology mapping should be done

when the standard ontology will be available; but, to

avoid of organizing the data relevant for user's

activity modelling by means of a folk ontology that

will require a great mapping effort when the

standard ontology will be available, in our

implementation we structured these data taking into

account the extended General User Model Ontology

(GUMO) outlined in fig.5, that is a meaningful step

towards a standard user model ontology.

Let us note that the ontology of fig.5 differs from

GUMO since it contains further data sections, e.g.,

10 cm

1 cm

Arduino

Controller

GPRS-

GSM

Shield

PECCS2014-InternationalConferenceonPervasiveandEmbeddedComputingandCommunicationSystems

the ones related to the car traffic and to the services

of city interest. Also, the terminology to represent

the data within each section takes into account the

one proposed by GUMO, as well as the terminology

proposed by a companion GUMO ontology called

UbisWorld (Heckmann, 2006). A reconciliation of

terms and properties should be done in future

works.

Since the paper deals with ambient data, in fig. 6

we point out how they are represented in Wi-City

and some differences with UbisWorld. To this aim,

on the top of fig.6 we sketch, by means of the

OntoGraph option of Protégé, the part of the Wi-

City ontology representing transport facilities and

urban services. Circles and rhombi denote

respectively classes and individual objects.

Figure 5: Extending the GUMO structure

Let us note that any ontology may be represented by

an object oriented (OO) language, but that an OO

program is not necessarily developed according to a

standard data ontology since object names and

properties may be chosen freely by the programmer.

Also, data ontology refers to the standard

terminology to be used in a certain knowledge

domain and may be more conveniently represented

by markup languages such as XML, RDF or OWL.

Therefore, differences between ontology

representations deal with the terms used and their

Figure 6: Some transport concepts used in Wi-City (on the

top) versus UbisWorld (on the bottom).

interrelations. Fig. 6 points out that in Wi-City the

name of a street (service) is an individual object

rather than one of the data fields of a street (service)

as suggested by UbisWorld (see the bottom of

fig.6). Also, in Wi-City the urban graph is defined as

a set of street traits delimited by crossroads,

whereas this aspect is not handled by UbisWorld.

The physiological conditions are well

represented in GUMO by the data collected by the

Biometrical Sensors, as pointed out in fig.5, e.g.,

EOG, EOM, and ACC sensors. But, in GUMO it is

not present a section dealing with the Environmental

Sensors, although such sensors should be foreseen

explicitly in the ontology to represent the things of

internet used to measure the indoor and outdoor

environment conditions.

Indeed, in GUMO the environment data are

described generically by variables referred to either

indoor or outdoor locations and collected under the

section Physical environment in fig.5.

Moreover, both in GUMO and UbisWorld the

properties of the ambient ontological entities, i.e.,

the ones linking the main subjects to the objects of

the ambient domain, are not defined explicitly.

Therefore, how linking the subjects to relevant

ambient objects by agreed properties (also named

predicates) is still an open problem.

A simple way to solve this problem is the one of

linking any subject of our domain, e.g., a person or a

car, to the location by the predicate IsLocatedIn and

Stores

EtneaDQ

EtneaQV

EtneaDS

Pharmacy

Supermarket

Crossroad

Street

Resource

BusStop

Zacco

Auchan

Barriera

Thing

UbisWorld

Street

Segments

Street

Location

atial

Elements

Etnea

Street

Europe

square

Personal

Demo

hic

Abilities

siolo

BioSensors

(

ECG

EOG

EOM

ACC

)

Current State

Location

User Preferences

Physical Environmen

has physiological data

is measured by

is located in

is featured by

Temperature

Pressure

Humidity

Luminosity

Preferred Media

Contents, Job,

Hobby, Holiday

Traffic Network

Car flows, Travelling

times, Road accidents,

Road works, ...

Cit

Services

Hospitals, Parks,

Pharms, Fuel Stations,

Public Offices, ...

General User Model Ontology (GUMO)

PervasiveAmbientIntelligencePlatformsintheIOTErabasedonaUbiquitousUserModelOntology-AnImplementation

Account

the location to the environment by predicates such as

ItsTemperatureIs or ItsHumidityIs. The values of a

variable should not be represented as individual

objects unless such values refer to a stable entity,

e.g., the name of a street may be seen as an

individual object whose temperature is given by a

data measured by a thermometer and the current

temperature value is a field of such data.

Following such conceptualization, the main

environmental data featuring the location of the

person named "Annalisa" may be represented by the

simple ontology shown in fig. 7 that should be read

as follows: Annalisa (subject/individual object) is

located in (predicate) Etnea Street (subject/

individual object) where there is the thermometer

T01 (subject/ individual object). The street

temperature is the data measured by T01, whose

current value is the current street temperature.

Analogously, the user heartbeats are not an

individual object but are data associated to the ECG

sensor worn by the user. This is represented in the

ontology of fig.8 where we take into account both

physiological and environmental ambient data.

Such ontology may be described as follows:

Annalisa is located in the Etnea Street and wears an

ECG, i.e., one of the wearable biometrical sensors.

In the street there is a temperature sensor that is one

of the sensors of a weather station. Her heartbeats

and the street temperature are given by the values

associated to the data heartbeats and temperature of

the ECG identified by ECG01 and by the

thermometer identified by T01.

Figure 7: Ontology describing some environment data

Figure 8: Ontology describing some ambient data at

different abstraction levels.

Although we have discussed only some aspects of

how building a suitable ambient ontology, it is easy

to understand that defining a well structured

ontology is a complex task, but once the ontology is

defined, it is simple to insert novel instances, such as

further monitoring devices, as soon as they are

installed in the system. For example, a novel ECG,

i.e., the one named ECG02 worn by John whose

current heartbeats are 75 bpm, may be represented

by inserting into the OWL file of the outlined

ontology the following few statements:

</ClassAssertion>

</ObjectPropertyAssertion>

<Literal datatypeIRI="http://www.w3.org/2001/

XMLSchema#float">75.0</Literal>

</DataPropertyAssertion>

Let us note that the last assertion deals with the ECG

measurement at a certain moment in time, and

therefore it should be stored into a data base with

date and day time information. In fact, at the

reception of an update ECG measurement the older

values are substituted by the new ones and are no

longer available on line.

The management of the data collected by the

environment sensors may be carried out analogously

as shown by the interface implemented at the Wi-

City server (fig.9): the current data coming from the

weather stations are on line, the older ones are

archived and may be analyzed by choosing the

option "archived data".

Figure 9: Wi-City Monitoring system to check the wheater

conditions and to analyze the archived data.

Since both the current and the older data are stored

in OWL format, any mobile may visualize the

Temperature

WEATHERSTATIONN.10

ViaNazionalec/oCannizzaroHospital

Street

Environment

SensorData

ermometer

T01

Etnea

T01IsInEtneaStreet

PersonAnnalisaThing

Sensor

Dimensions

Environment

SensorData

ACC

ECG

Temperature

Biometrical

SensorData

T01

ECG01

Person

Annalisa

Street

Etnea

Street

T01temperature25°C

Archived

DATA

EOG

EMG

Temperature BMP

Temperature Grove

Weather Condition

Wind Speed

Rain Quantity

Pressure BMP

Humidity Grove

Wind Direction

PECCS2014-InternationalConferenceonPervasiveandEmbeddedComputingandCommunicationSystems

current and previous ambient data on its display as

shown in fig.10 independently on how the data were

collected by the proprietary monitoring systems.

Figure 10: User interface to display on the mobile the road

weather conditions, map and photo in real time.

4 CONCLUSIONS

In the paper we have illustrated how the

implementation of a pervasive ambient intelligence

platform in the IOT era should be based on a

ubiquitous user's model ontology. Indeed, the

availability of cheap and small monitoring devices

addressable through internet is only one side of the

coin, since it is also important that the monitored

data should be open and interoperable.

This clarifies why the dedicated navigators

installed on the cars that don’t take into account the

real time car traffic flows neither the personal and

weather conditions are more and more substituted by

modern applications implemented on the user

mobiles that not only facilitate e-commerce and e-

government operations but also help the people

mobility using timely information coming from field

sensors, as foreseen in (TRG, 2008). In fact, such

data are able to characterize the user status in a

deepened way and the traffic and weather conditions

in real time, thus allowing the DSS to satisfy

effectively the ubiquitous request of user assistance.

However, the available Location Based Services

(LBSs) of this second generation are mainly

proprietary systems, thus they don't meet the basic

requirements of the LBSs inspired by the IOT

paradigm, i.e., the requirement that the data of user

interest should be open and interoperable, as claimed

in (Teller, 2010).

For this reason, the ontology approach until now

used in K-Metropolis for integrating disparate urban

data bases to support user mobility and to provide e-

commerce and e-government services to desktop

PCs and mobiles, was extended, as illustrated in this

paper, to make available to all the DSSs

implemented on the user mobiles the data needed to

provide the users with recommendations that take

into account personal and ambient information that

influence greatly the user activities.

However, due to the lack of an agreed ontology

at urban level, our future work will be mainly the

one to study carefully the available urban ontology,

e.g., (Heckmann et al., 2005) and (Heckmann,

2006), (Teller, 2007), (Berdier, 2007), (Zhai, 2008),

(Faro, 2011b) and (Costanzo, 2013b), to choose the

terminology and related properties that may favour

the implementation of a standard smart city

ontology.

REFERENCES

Acampora G., et al., A Survey on Ambient Intelligence in

Health Care, IEEE Proceedings, Issue 99, 2013.

Aoun C., The Smart City Cornerstone: Urban Efficiency.

Schneider Electric White Paper, 2013.

Banzi, M., Getting Started with Arduino. Make Books,

2009.

Berdier C., Roussey C., Urban Ontologies: the Towntology

Prototype towards Case Studies. Springer, 2007.

Berthon B., et al., Building and Managing an Intelligent

City, online, 2011.

Corlan, M., Flash Platform Tooling: Flash Builder, Adobe,

2009.

Costanzo A., Faro A., Giordano D., Wi-City: Living,

deciding and planning using mobiles in intelligent

cities, 3rd Int. Conf. on Pervasive Embedded

Computing and Communication Systems, PECCS,

Barcelona, 2013.

Costanzo A., Faro A., Towards Open and Interoperable

Information Platforms for Mobile Users: RDF

Methodology and Case Study, IEEE Proc. 7th

International Conf. on Application of Information and

Communication Technologies (AICT), Baku, 2013.

Costanzo A., Pino C., An Arduino based Health

Monitoring System for Elderly People Living at

Home: Functions and Ontology, Int. Conf. on

Physiological Computing Systems, PhyCS, INSTICC,

Lisbon, 2014.

Faro, A., Giordano, D. and Spampinato, C., Evaluation of

the Traffic Parameters in a Metropolitan Area by

Fusing Visual Perceptions and CNN Processing of

Webcam Images, IEEE Transactions on Neural

Networks, Vol.19 (6), 1108-1129, IEEE, 2008.

Faro, A., Giordano, D. and Spampinato, C., Adaptive

background modeling integrated with luminosity

sensors and occlusion processing for reliable vehicle

detection. IEEE Transactions on Intelligent

Transportation Systems. Vol.12(4), 1398-1412, IEEE,

2011.

ViaNazionalec/o

CannizzaroHospital

WEATHERSTATIONN.10

PervasiveAmbientIntelligencePlatformsintheIOTErabasedonaUbiquitousUserModelOntology-AnImplementation

Account

Faro A., Giordano D., Spampinato C., Integrating

Location Tracking, Traffic Monitoring and Semantics

in a Layered ITS Architecture. Intelligent Transport

Systems, vol.5(3),IET, 2011.

Heckmann D., et al., GUMO, the General User Model

Ontology, LNCS, N.3538, Springer, 2005.

Heckmann D., Ubiquitous User Modeling, IOS Press,

2006.

Leduc, G., Road Traffic Data: Collection Methods and

Applications, Working Papers on Energy, Transport

and Climate Change, N.1, JRC European

Commission, 47967, 2008.

Telematics Reserach Group (TRG), Preview portable

navigation: the future is bright for connectivity.

http://www. telematicsresearch.com/PDFs/

TRG_Press_ Jan_0.pdf.2008.

Teller J., John R. Lee J.R., Catherine Roussey C., (Eds.),

Ontologies for Urban Development, Springer, 2007.

Teller J., Billen R., Cutting-Decelle A.F., Ontology based

approaches for improving the interoperability between

3D urban models. Journal of Inf. Technology in

Construction, 2010.

Zhai, J., Jiang, J., Yu, Y., Li J., Ontology-based Integrated

Information Platform for Digital City. IEEE Proc. of

Wireless Communications, Networking and Mobile

Comp., WiCOM ’08, 2008.

PECCS2014-InternationalConferenceonPervasiveandEmbeddedComputingandCommunicationSystems