IoT Semantic Interoperability: A Systematic Mapping Study

Amanda D. P. Venceslau

, Rossana M. C. Andrade

, Vânia M. P. Vidal

, Tales P. Nogueira

and Valéria M. Pequeno

Federal University of Ceará, Campus de Crateús, Crateús-CE, Brazil

Department of Computing, Federal University of Ceará, Campus do Pici, Fortaleza-CE, Brazil

Group of Computer Networks, Software Engineering and Systems (GREat), Federal University of Ceará,

Campus do Pici, Fortaleza-CE, Brazil

TechLab, Departamento de Ciências e Tecnologias, Universidade Autónoma de Lisboa Luís de Camões, Portugal

Keywords:

Semantic Interoperability, IoT Semantic Interoperability, Semantic Web, Internet of Things, Ontology.

Abstract:

The Internet of Things (IoT) is a paradigm in which the Internet connects people and the environment using

devices and services that are spread in the user daily routine. In this scenario, different agents, devices and

services are able to exchange data and knowledge using a common vocabulary or mappings that represent

and integrate heterogeneous sources. This semantic interoperability is facilitated by the Semantic Web that

provides consolidated technologies, languages and standards, offering data and platforms interoperability. In

this context, this work reviews and analyzes the state-of-the-art of IoT semantic interoperability, investigating

and presenting not only which Semantic Web technologies are employed but also the challenges that support

the studies in this area of research.

1 INTRODUCTION

The Internet of the Things (IoT) emerged as a

paradigm in which people and things exchange infor-

mation anytime, anywhere, with any device. In the

IoT environment, sensors are able to monitor people

and the environment, generating a large volume of

data that can be consumed by different applications

in different domains, e.g., cardiac monitoring, smart

homes, and smart cities management.

Almost a hundred new IoT platforms entered the

global market in 2016, increasing the total number

to about 450 (Ganzha et al., 2017b). It is forecasted

that about 50 billion devices will be connected to the

Internet by 2020 (Evans, 2011). In this scenario of

increasing number of distributed and heterogeneous

devices, there is a demand for technologies that sup-

port interoperability, providing means of representa-

tion, discovery and integration.

The use of higher-level abstractions, i.e., seman-

tic information extracted from data, is considered an

attractive solution to minimize heterogeneity employ-

ing common and shared representations. In this con-

text, semantic interoperability can be deﬁned as the

ability of different agents, services, and applications

to exchange information, data and knowledge both on

and off the Web (W3C, 2018).

Within IoT environments, where data are hetero-

geneous, dynamic and distributed, mechanisms of se-

mantic interoperability are necessary to provide inter-

operability in a ﬂexible way, allowing systems to un-

derstand the collected data. Through Semantic Web

standards, semantic interoperability agents can use

shared vocabularies to exchange information between

platforms (W3C, 2018).

The European Research Cluster on the Internet of

Things (IERC) released a set of best practices and rec-

ommendations for semantic interoperability (Serrano

et al., 2015). However, this study examines the chal-

lenges encountered in IoT semantic interoperability,

but does not show which Semantic Web technologies

have been adopted and what are the limitations of cur-

rent approaches.

Rigorous research that collect and present the

technologies and challenges related to semantic inter-

operability in the context of IoT has not been found

in the literature so far. In this perspective, this pa-

per presents a systematic mapping study (Kitchen-

ham et al., 2010) about semantic interoperability in

the context of IoT with the purpose of deﬁning the

state-of-the-art, emphasizing which and how Seman-

tic Web technologies have been used and what are the

Venceslau, A., Andrade, R., Vidal, V., Nogueira, T. and Pequeno, V.

IoT Semantic Interoperability: A Systematic Mapping Study.

DOI: 10.5220/0007732605350544

In Proceedings of the 21st International Conference on Enterprise Information Systems (ICEIS 2019), pages 535-544

ISBN: 978-989-758-372-8

535

current limitations and challenges faced in the ﬁeld.

The remainder of this paper is structured as fol-

lows: Section 2 provides the background for the pa-

per with a brief explanation about Semantic Interoper-

ability and IoT Semantic Interoperability; in Section

3, the process of systematic mapping study applied

to this work is described; in Section 4, the results

obtained from the systematic mapping study are pre-

sented and analyzed based on the research questions;

Section 5 addresses the study limitations; and, in Sec-

tion 6, we discuss the ﬁndings that emerged from the

results and present our ﬁnal considerations.

2 IOT SEMANTIC

INTEROPERABILITY

Semantic Interoperability refers to the ability of two

or more computational systems to exchange informa-

tion through a shared meaning that can be interpreted

automatically and correctly. Thus, interoperability at

the semantic level requires a common understanding

of the meaning of the content being exchanged, pre-

serving the semantics of the original message.

Despite the adoption of shared ontologies and vo-

cabularies that allow the representation and sharing

of meaning, there is still a need to maintain the intrin-

sic information from data sources, preserving domain

knowledge and facilitating data maintenance. An-

other issue of semantic interoperability is related to

the maintenance of shared ontologies, which requires

a centralized and periodic updating mechanism. Fi-

nally, the software infrastructure based on shared on-

tologies still has problems related to scalability, a

well-known open question of the Semantic Web.

The data generated by IoT devices in different for-

mats hinders the interoperability of applications and

platforms that can not interpret the data, acting incon-

sistently on the received information. In this context,

there is a need for using common vocabularies capa-

ble of describing the meaning of data in this environ-

ment. Semantic interoperability is a concept only re-

cently studied in the context of the Semantic Web. It

has gained prominence in both academia and indus-

try, that have been applying its principles to IoT sce-

narios (Gyrard et al., 2018). In order to implement

solutions that minimize the interoperability problems

found in the IoT environment, the community began

to adopt Semantic Web technologies, standards, lan-

guages, and approaches for modeling (Serrano et al.,

2015) and integrating ontologies.

Semantic Web technologies provide the technical

and operational structure as well as the means to facil-

itate semantic interoperability. This can be achieved

by either modeling new ontologies or reusing existing

ontologies by semantically integrating (aligning) dif-

ferent vocabularies with equivalences between their

classes and properties (W3C, 2018).

However, the use of these technologies still faces

challenges in terms of sharing the published ontolo-

gies and correct and consistent reuse. More specif-

ically, there are several requirements and challenges

for semantic interoperability in IoT that should be

addressed by other tools, for instance: integration

of distributed data sources, a uniﬁed semantic an-

notation model of IoT data, management of sensors

based on composition and fusion of streams from

various data sources, discovery of sensors and data

sources for application requests according to their ca-

pabilities, analysis and reasoning on semantic level

resources through reasoners and visualization tools

(Serrano et al., 2015).

In spite of this effort to elicit the limitations of

semantic interoperability approaches and to describe

auxiliary tools, IERC does not recommended Seman-

tic Web approaches, tools or methodologies to: (i)

Semantic Web methodologies to model IoT ontolo-

gies, (ii) reuse of existing ontologies and (iii) valida-

tion tools for ontologies (Gyrard et al., 2018).

Gyrard and Serrano (2015) proposed a uniﬁed se-

mantic mechanism, presenting technologies for the

construction of interoperable systems in the domain

of IoT and Smart Cities. One of the main chal-

lenges highlighted by the authors is to unify mod-

els, vocabularies and ontologies to semantically anno-

tate the data. The authors cite the LOV4IoT

catalog

of ontologies and the need for approaches and tools

that provide a uniﬁed semantic model aligned with

the existing vocabularies provided by IoT platforms.

The authors extend the M3 ontology (Gyrard et al.,

2015a), which is already aligned with ontologies pro-

vided by IoT platforms, to address the weaknesses re-

lated to ontologies (Gyrard and Serrano, 2015). How-

ever, their work does not discuss limitations related

to best practices for ontology modeling, reuse and

matching.

These and other limitations of current approaches

encourage work in this area to develop studies that ad-

dress these gaps, pointing out tools and solutions that

drive the adoption of semantic interoperability mech-

anisms in the IoT scenario.

http://lov4iot.appspot.com

ICEIS 2019 - 21st International Conference on Enterprise Information Systems

536

3 THE RESEARCH METHOD:

SYSTEMATIC MAPPING

STUDY

A systematic mapping study (Kitchenham et al.,

2010) is a means to identify, evaluate and interpret rel-

evant work concerning a research question or a phe-

nomenon of particular interest. It is assumed that

the results of using a controlled process and a for-

mal bibliographic research beneﬁt the research com-

munity by listing topics of interest, gaps, and chal-

lenges. Associated to our systematic mapping study,

we also performed snowballing procedures (Wohlin,

2014), a systematic research approach that uses back-

ward (i.e., checking the references of the studies) and

forward (i.e., checking papers that cited the studies)

to identify additional papers, in order to expand the

set of works considered in our research.

The research questions for a systematic mapping

study are more general, including questions of which

sub topics were addressed and which sub topics have

sufﬁcient studies for detailed review.

We have followed a process based on guide-

lines for performing systematic literature reviews by

Kitchenham and Charters (2007) that is shown in

Fig 1. The process consists of three main activities

that are detailed in the following sections: planning

(Section 3.1), conducting (Section 3.2), and reporting

(Section 4) . The planning activity aims to deﬁne the

protocol, organizing the research steps. In our work,

the conducting activity was executed in two distinct

phases. First, we selected the primary studies using

the digital libraries. After, we complemented the set

of articles with the snowballing procedure (Wohlin,

2014).

3.1 Planning: Protocol Deﬁnition

The aim of the planning phase is the deﬁnition of

a review protocol. The systematic review protocol

(Kitchenham, 2004) deﬁnes the research questions to

be answered, the sources, and how papers are se-

lected. Thus, for the execution of this work, a pro-

tocol was developed. Its main topics are described in

the following.

3.1.1 Research Questions

The objective of our study was to identify the Seman-

tic Web technologies and the challenges to ensure se-

mantic interoperability in the context of IoT. There-

fore, we have established the following research ques-

tions:

RQ1: What Semantic Web technologies have been

proposed to ensure semantic interoperability in the

context of IoT?

RQ2: What limitations and challenges to ensure se-

mantic interoperability in the context of IoT have

been described by these proposals?

3.1.2 Search String

Based on the PICO (P - patient, problem or popula-

tion; I - intervention; C - comparison, control or com-

parator; O - outcome) approach (Pai et al., 2004) and

on the key terms of the research ﬁeld, the following

search string was elaborated: ((“internet of things”

OR “iot” OR “web of things” OR “wot”) AND (“se-

mantic interoperability” OR “semantic integration”

OR “ontology integration”) AND (“state of the art”

OR survey OR problem OR “lessons learned” OR

middleware OR challenge OR application)).

The term semantic integration was added as part

of a broader view of semantic interoperability as,

driven by semantic integration, services and tools can

become interoperable (W3C, 2018).

3.1.3 Research Sources

To obtain the primary studies, we have used two

methods: database search and snowballing. For the

database search, we selected the most relevant digital

libraries in Computer Science and Engineering (Chen

et al., 2010): (a) ACM Digital Library

; (b) IEEE

Xplore Digital Library

; and (c) Science Direct

. Re-

garding the snowballing approach, we used both the

backward and forward procedures (Wohlin, 2014).

3.1.4 Study Selection Criteria

We have deﬁned the following selection criteria in or-

der to select the most suitable studies:

SC1: The study must be written in English;

SC2: The selected article must be available on the

Web;

SC3: The study should present initiatives related to

semantic interoperability in the IoT domain, covering

at least one of the research questions.

No restriction was deﬁned based on the type of

works, which means that all kinds of studies, such as

conference and journal papers, short and full papers,

were chosen. They were processed in the same way

considering the above selection criteria. There is also

no restriction on the publication years of the papers.

http://dl.acm.org

https://ieeexplore.ieee.org

https://www.sciencedirect.com

IoT Semantic Interoperability: A Systematic Mapping Study

537

Figure 1: The systematic mapping process.

Table 1: Number of papers eliminated by selection criteria.

Selection criteria Number of papers

SC1 0

SC2 13

SC3 - research question (RQ1) 134

SC3 - research question (RQ2) 10

3.2 Conducting

This activity was performed in two phases. The ﬁrst

phase is related to the conduction of database search,

which aims to ﬁnd relevant papers in digital libraries

using well-deﬁned search strings. The second phase

is related to the conduction of backward snowballing,

i.e., seeking papers from reference lists of the papers

identiﬁed during the ﬁrst phase, and the conduction

of forward snowballing, i.e., seeking papers that had

cited the papers found during the ﬁrst phase of the ac-

tivity. The outcome of the conducting activity phases

are described in the following.

3.2.1 Conducting: First Phase

In this step, we searched for articles based on the de-

ﬁned protocol. The set of search strings was applied

to the search engines: ACM, IEEE and Science Di-

rect. This stage retrieved 169 papers, 36 from the

IEEE, 13 from the ACM and 120 from Science Di-

rect. After the application of the selection criteria, we

rejected 157 papers (shown in Tab. 1) and 12 papers

remained. Finally, we performed data extraction for

each selected paper, extracting approaches, technolo-

gies and tools aimed at the modeling, reuse and vali-

dation of ontologies and the challenges in their adop-

tion.

3.2.2 Conducting: Second Phase

The snowballing procedure is usually performed with

a starting set of papers. In our study, the starting

set corresponded to the 12 papers selected in the ﬁrst

phase of the conducting activity. Moreover, we per-

formed backward and forward snowballing to ﬁnd ad-

ditional papers. In the end, we obtained four more

articles for data extraction.

4 RESULTS OF THE

SYSTEMATIC MAPPING

STUDY

This systematic mapping study was performed in De-

cember, 2018. We have found 16 papers (shown in

Tab. 2) that answer the deﬁned research questions. All

of them present approaches that use Semantic Web

technologies and discuss limitations and challenges

to ensure IoT Semantic Interoperability. Only two pa-

pers (Ganzha et al., 2017b,c) propose semantic inte-

gration approaches, presenting steps such as a map-

ping among ontologies.

Moreover, these papers cover different domains

of application such as health, transport and logistics

(Ganzha et al., 2016), geospatial data (Ganzha et al.,

2017c) and water management (Howell et al., 2018).

The results of this systematic study are discussed in

details in Section 4.1, which discusses the results re-

lated to the Semantic Web technologies proposed for

IoT Semantic Interoperability and proposes a list of

methods and tools suitable for IoT Semantic Interop-

erability, and Section 4.2, which presents the limita-

tions and challenges to ensure IoT Semantic Interop-

erability found in the proposed approaches.

4.1 Semantic Web Technologies

Proposed for IoT Semantic

Interoperability

Semantic Web technologies provide knowledge shar-

ing and reuse mechanisms (Simperl, 2009), which fa-

cilitate semantic interoperability. However, the use

of Semantic Web technologies for semantic interoper-

ability approaches needs to follow the Semantic Web

community recommendations in order to boost the

reuse of existing knowledge, avoiding the construc-

tion of heterogeneous models that hinder interoper-

ICEIS 2019 - 21st International Conference on Enterprise Information Systems

538

Table 2: IoT Semantic Interoperability approaches.

Aspects Approaches

Model ontologies

Gyrard et al. (2015b), Agarwal et al. (2016), Gyrard and Serrano (2016),

Strassner and Diab (2016), Gyrard and Serrano (2015), Ganzha et al. (2017a) ,

Howell et al. (2018)

Reuse ontologies

Gyrard et al. (2018), Gyrard et al. (2015b), Gyrard and Serrano (2016),

Gyrard et al. (2018), Strassner and Diab (2016), Rhayem et al. (2017),

Ganzha et al. (2016), Barnaghi et al. (2012), Ganzha et al. (2017b),

Ganzha et al. (2017c),Gyrard and Serrano (2015),Chindenga et al. (2016)

Nagib and Hamza (2016)

Ontology validation tools Gyrard et al. (2018)

ability. Semantic interoperability also refers to other

levels and tasks, such as semantic mappings and rea-

soning. However, these issues were not raised by the

articles found during this study.

In order to reply RQ1, during the data extraction

process, we identiﬁed approaches and tools that: (i)

Semantic Web methodologies to model IoT ontolo-

gies, (ii) reuse existing ontologies, and (iii) ontology

validation tools.

4.1.1 Semantic Web Methodologies to Model

IoT Ontologies

Following the best practices discussed by Gyrard

et al. (2015b), we found methodologies that encour-

age the modeling and reuse of ontologies, for ex-

ample, the ones by Noy et al. (2001) and Suárez-

Figueroa (2010), called NeON, which provide tutori-

als with guidelines for developing well-designed on-

tologies. The work of Noy et al. (2001) is also used

as a methodology for ontology construction by Agar-

wal et al. (2016), which also proposes an ontology

available online, through a open source ontology doc-

umentation tool called Live OWL Documentation En-

vironment (LODE)

The proposal of Gyrard and Serrano (2016)

presents a methodology for semantic interoperability

applied to IoT and smart cities that adapts the NeON

methodology to speciﬁc characteristics of IoT. How-

ell et al. (2018) applied the NeON methodology and

adapted the existing semantic resources to the appli-

cation of IoT aimed at the domain of water manage-

ment. One of the goals of their work is to develop a

reference ontology in the ﬁeld.

There is a need to encourage best practices by de-

veloping ontologies (Gyrard et al., 2018), reusing ex-

isting ontologies as much as possible and aligning on-

tologies to increase interoperability, reducing hetero-

geneity between models and development time.

As discussed by Gyrard and Serrano (2015), lim-

itations such as the lack of adoption of best practices

http://www.essepuntato.it/lode

can be solved with approaches presented here, such

as the one of Noy et al. (2001) and NeON (Suárez-

Figueroa, 2010), in order to deﬁne a uniﬁed model

for modeling ontologies in IoT, see Fig 2.

4.1.2 Reuse Existing Ontologies

The ﬁrst step in the reuse of existing ontologies

is through catalogs (Gyrard et al., 2018) such as

the Linked Open Vocabulary for Internet of Things

(LOV4IoT), which referenced 448 ontologies relevant

to IoT in May 2018 (Gyrard et al., 2015b; Gyrard

et al., 2018). It includes a status ﬁeld that indicates

whether the ontology is shared online or follows the

best practices, which is the main difference to other

catalogs. As a catalog of ontology sharing and reuse,

proposal (Gyrard and Serrano, 2016) also adopts the

LOV4IoT.

There are other catalogs that aimed at the con-

text of smart cities, for instance, Ready4Smartcities

(Poveda-Villalón et al., 2014a), a catalog that in-

tegrates the OOPS (Poveda-Villalón et al., 2014b)

ontology validation tool and the OpenSensingC-

ity

catalog, which also provides ontologies in the

context of smart cities, including the WebVOWL

(Lohmann et al., 2014) visualization tool, OOPS and

the TripleChecker

tool for ontology syntax valida-

tion (Gyrard et al., 2018).

The proposals of Gyrard et al. (2018); Gyrard

et al. (2018) investigated relevants IoT catalogs, i.e.,

Ready4Smartcities, OpenSensingCity and LOV4IoT,

presenting a methodology for enhancing catalogs of

ontologies, supporting their maintenance.

Among the ontologies found in these catalogs the

W3C Semantic Sensor Network (SSN) ontology is

recognized as the standard for generically describing

information in the IoT environment and some works

have extended this ontology to applications of speciﬁc

domains (Gyrard et al., 2015b; Rhayem et al., 2017).

Any project that uses semantic technologies in IoT

http://opensensingcity.emse.fr

http://graphite.ecs.soton.ac.uk/checker

IoT Semantic Interoperability: A Systematic Mapping Study

539

should extend the SSN ontology, adding the concepts

needed to handle the application in the intended do-

main (Ganzha et al., 2016; Barnaghi et al., 2012). The

choice of a consolidated ontology like SSN avoids the

use of a set of ontologies that describe the same con-

cepts (Strassner and Diab, 2016).

To facilitate the exchange of information be-

tween two or more IoT artifacts, ontological align-

ments (Ganzha et al., 2017b) can be used to trans-

late messages between entities with different seman-

tic representations of the domain of interest. Align-

ment of ontologies refers to the process of ﬁnding

matches between two or more ontologies. To repre-

sent this alignment the INTER-IoT project develops

its own INTER-IoT Alignment Format (Ganzha et al.,

2017b,c), an XML representation inspired by the API

Alignment (David et al., 2011) and to some extent by

EDOAL

Gyrard and Serrano (2015) have shown the im-

portance of a catalog of ontologies (LOV4IoT in this

case) to support the proposed uniﬁed semantic mech-

anism. Other catalogs for smart cities, described in

this section, are important in this approach, such as

Ready4Smartcities and OpenSensingCity. In addi-

tion, the authors cite the lack of ontology matching

or alignment tools adapted for IoT ontologies. In this

section, we described an approach that promises to

overcome this limitation, called INTER-IoT Align-

ment Format, see Fig 2.

4.1.3 Ontology Validation Tools

To ensure that ontologies comply with best practices,

it is necessary to introduce tools that validate its syn-

tax, detecting undeclared elements or made available

in an incorrect format (Gyrard et al., 2015b). The

work of Gyrard et al. (2018) performed an evaluation

of 27 ontologies for IoT and smart cities, available

in the LOV4IoT catalog, using six validation tools,

namely: Parrot

, WebVOWL, Oops, TripleChecker,

LODE and Vapour

. This evaluation concluded that

there are ontologies that can not be loaded in all tools,

revealing errors, suggesting LODE as preferred tool

in comparison to Parrot due to the possibility of auto-

matic documentation of more ontologies.

4.2 Limitations and Challenges

In order to answer RQ2, during the data extraction

process, we extracted the limitations and challenges

http://alignapi.gforge.inria.fr/edoal.html

http://ontorule-project.eu/parrot/parrot

http://linkeddata.uriburner.com:8000/vapour

that the existing proposals face in order to ensure se-

mantic interoperability in the context of IoT.

4.2.1 Semantic Web Methodologies to Model

IoT Ontologies

Initially, IoT data and knowledge engineering ef-

forts focused on developing IoT data infrastructure by

means of publication and data access, giving less at-

tention to data processing and integration with exist-

ing applications (Barnaghi et al., 2012).

Inside the IoT community, each project, platform

or application usually develops its own ontology (Gy-

rard and Serrano, 2015). Few studies use a method-

ology for ontology modeling (Agarwal et al., 2016;

Gyrard and Serrano, 2016), making interoperability

difﬁcult and reducing the possibility of reuse (Gyrard

et al., 2015b).

The Semantic Web community uses practices

(Gyrard et al., 2014) that have not been followed in

the IoT scenario (Gyrard and Serrano, 2015). It is

noticed the absence of Semantic Web experts in the

modeling of the ontologies found for this scenario.

According to Ganzha et al. (2017a) the organizations

that develop ontologies modeling approaches do not

seek consensus, which is a social aspect that limits

interoperability in the IoT environment and becomes

a challenge.

Another aspect is related to the language and for-

malism in which the developed ontologies have been

made available. The approach presented by Ganzha

et al. (2017a) in the context of healthcare and logistics

and transport does not have an ontology represented

by Semantic Web languages such as RDF(S)

OWL

, impairing the sharing of ontologies and the

ability to exchange information between agents and

platforms. Nevertheless, ontologies can be written in

different formalisms, some more expressive than oth-

ers (Strassner and Diab, 2016). In domains like the

ones described by Strassner and Diab (2016), data in-

tegration from a common vocabulary is difﬁcult, be-

cause most available ontologies are either deﬁned in

UML artifacts or described in markup languages. It

is required to normalize concepts between formal on-

tologies and implicit ontologies, making it a challenge

for non-specialist users.

4.2.2 Reuse Existing Ontologies

Some approaches present the use of ontologies in the

IoT scenario (Agarwal et al., 2016), but they miss or

https://www.w3.org/RDF

http://www.w3.org/OWL

ICEIS 2019 - 21st International Conference on Enterprise Information Systems

540

have insufﬁcient concepts for the measurements pro-

vided by the sensors, and many of these ontologies

do not follow the best practices, thus, it is difﬁcult to

correctly interpret the concepts.

The reuse of existing standard domain ontologies

and higher ontologies is suitable to unify high-level

concepts in various applications and should be ex-

tended according to the intended application logic

(Nagib and Hamza, 2016). The SSN ontology

presents itself as a good example of a higher ontol-

ogy that contains classes focused on the concepts of

sensors and observations.

However, one of the problems of semantic inter-

operability (Amato et al., 2011) is related to the adop-

tion by many systems and approaches (Strassner and

Diab, 2016) of a collection of ontologies or a set of

data models to prototype an ontology that shares the

meaning of context in a common vocabulary, result-

ing in the loss of intrinsic information of the data

sources.

In the Semantic Web, this problem can be miti-

gated by using matching or alignment tools. However,

according to Gyrard and Serrano (2015), the seman-

tic level of class-to-class mapping used by ontology

matching or alignment tools is not enough to describe

the data in the IoT scenario, making it an open chal-

lenge.

Another aspect pointed out in Amato et al. (2011)

concerns the maintenance of shared ontologies. In

the IoT scenario, the catalogs of ontologies are used

to minimize this problem, however, the maintenance

of these catalogs is still a problem that deserves in-

vestigation (Gyrard et al., 2018; Gyrard et al., 2018).

Finally, scavenging is also a key issue in this en-

vironment, where, according to Chindenga et al.

(2016), practical mechanisms and implementations

are needed to provide independent interoperability of

manufacturers and suppliers of devices and services,

ensuring efﬁcient scalability of IoT.

5 THREATS TO VALIDITY

When conducting secondary studies, the ﬁndings

made by the researchers, both in the selection of the

studies and in the conclusions inﬂuence the result

(Wohlin et al., 2012). Thus, in this section we present

the threats to the validity of our study.

As threats to the validity of the results of the sys-

tematic mapping study carried out are: (i) bias in the

selection of the studies analyzed; (ii) imprecision in

https://www.w3.org/TR/vocab-ssn/

the data extraction and (iii) how to classify and inter-

pret data.

The selection of the analyzed studies followed the

selection procedure of the primary studies, in which

the search of the sources was applied, an initial set of

studies was selected from the titles and abstracts of all

papers. We also performed the snowballing process

from the initial set of selected studies. After com-

paring the papers using the inclusion and exclusion

criteria (Section 3.1), the selected papers were fully

read and, again, faced with the criteria. The included

articles were selected for data extraction.

In order to minimize the loss in data extraction,

the advanced mode of search engines was used in all

search sources. Papers that did not use the keywords

deﬁned in our search string in the title or abstract were

not selected. Hence, papers that did not mention IoT

were not listed in this paper. Finally, the data extrac-

tion was performed in pairs and, in the case of impass,

the discussions were done with a third reviewer until

a consensus was reached.

Regarding the data classiﬁcation, it was based on

limitations previously proposed in the literature, i.e.,

Semantic Web methodologies to model IoT ontolo-

gies, reuse of existing ontologies and validation tools

for ontologies. Data extraction and classiﬁcation were

performed by the ﬁrst author and reviewed by the

third author.

6 FINAL CONSIDERATIONS

This paper presented a systematic mapping study that

found 16 publications related to IoT Semantic Inter-

operability. These publications were analysed and we

discussed their limitations and challenges, and how

the existing technologies are relevant to the current

state-of-the-art.

Before performing the systematic mapping study,

we investigated the literature in search for secondary

studies on IoT semantic interoperability. Since no

secondary study was found on this research topic, we

performed the systematic mapping study. The map-

ping results provided an overview of the research re-

lated to the investigated topic.

Regarding Semantic Web technologies proposed

for IoT semantic interoperability, few approaches

were found for ontology modeling. The works that

adopted some approach needed to adapt some stages

of the process to IoT characteristics. We observed

that the lack of selection of the best practices rec-

ommended by the Semantic Web community is still

a gap that was not totally fulﬁlled by the existing ap-

proaches. With respect to the reuse of existing on-

IoT Semantic Interoperability: A Systematic Mapping Study

541

Figure 2: Approaches and limitations for a uniﬁed semantic engine for Internet of Things. Based on the work of Gyrard and

Serrano (2015).

tologies, there was a concern on the part of the IoT

community to build catalogs of ontologies aimed at

the domain of their applications as well as to adopt

ontology validation tools within the cataloging envi-

ronment, providing veriﬁcation of the ontology syn-

tax.

The approach of Gyrard and Serrano (2015) pre-

sented an aspect not yet explored by the IoT commu-

nity, which is the semantic level used by the current

matching or alignment tools, which, according to the

authors, are not enough to describe the data in the

IoT scenario. However, to minimize this problem, the

only approach that proposes a solution by developing

its own alignment format is the INTER-IoT Align-

ment Format (Ganzha et al., 2017b,c). Thus, we con-

clude that this is an unexplored research subject, as it

is important to maintain semantic interoperability in

this scenario.

As another result, we also ﬁnd approaches that

address other known semantic interoperability issues,

as the maintenance of intrinsic information the data

sources (Amato et al., 2011). The adoption of shared

ontologies solves the problem of single interpretation

of meaning, but may lead to the loss of relevant do-

main information from data sources.

Therefore, semantic interoperability in the con-

text of IoT has requirements that can be supported

by Semantic Web technologies, but there are limita-

tions in this context, especially due to formal repre-

sentation and correspondence of ontologies. Unfor-

tunately, these aspects, in general, can not be solved

automatically and, when we consider the literature in

this aspect, it is not clear how good the tools are in

order facilitate the work that must be performed.

In summary, these results indicate gaps in the

context of IoT semantic interoperability (i) need for

adopting best practices recommended by the Seman-

tic Web community; (ii) absence of methodologies in

modeling ontologies that meet the needs of applica-

tions in the domain of IoT; (iii) unavailability of main-

tenance of ontology catalogs in IoT domains; and (iv)

lack of matching or alignment tools to describe the

data in the IoT context. These gaps provide a roadmap

of issues to be explored in future research.

ACKNOWLEDGMENTS

This work is partially supported by INES

(www.ines.org.br), CNPq grant 465614/2014-0 and

FACEPE grants APQ-0399-1.03/17 and PRONEX

APQ/0388-1.03/14.

REFERENCES

Agarwal, R., Fernandez, D. G., Elsaleh, T., Gyrard, A.,

Lanza, J., Sanchez, L., Georgantas, N., and Issarny,

V. (2016). Uniﬁed IoT ontology to enable interoper-

ability and federation of testbeds. In Internet of Things

(WF-IoT), 2016 IEEE 3rd World Forum on, pages 70–

75. IEEE.

Amato, F., Fasolino, A. R., Mazzeo, A., Moscato, V., Pi-

cariello, A., Romano, S., and Tramontana, P. (2011).

ICEIS 2019 - 21st International Conference on Enterprise Information Systems

542

Ensuring semantic interoperability for e-health appli-

cations. In Complex, Intelligent and Software Inten-

sive Systems (CISIS), 2011 International Conference

on, pages 315–320. IEEE.

Barnaghi, P., Wang, W., Henson, C., and Taylor, K. (2012).

Semantics for the internet of things: early progress

and back to the future. International Journal on

Semantic Web and Information Systems (IJSWIS),

8(1):1–21.

Chen, L., Babar, M. A., and Zhang, H. (2010). Towards

an evidence-based understanding of electronic data

sources. In Proceedings of the 14th International Con-

ference on Evaluation and Assessment in Software En-

gineering, EASE’10, pages 135–138, Swindon, UK.

BCS Learning & Development Ltd.

Chindenga, E., Gurajena, C., and Thinyane, M. (2016). To-

wards an adaptive ontology based model for interoper-

ability in Internet of Things (IoT). In IST-Africa Week

Conference, 2016, pages 1–8. IEEE.

David, J., Euzenat, J., Scharffe, F., and Trojahn dos Santos,

C. (2011). The Alignment API 4.0. Semant. web,

2(1):3–10.

Evans, D. (2011). The Internet of Things: How the

next evolution of the Internet is changing everything.

CISCO white paper, 1(2011):1–11.

Ganzha, M., Paprzycki, M., Pawlowski, W., Szmeja, P., and

Wasielewska, K. (2016). Semantic Technologies for

the IoT - An Inter-IoT Perspective. In 2016 IEEE

First International Conference on Internet-of-Things

Design and Implementation (IoTDI), pages 271–276.

Ganzha, M., Paprzycki, M., Pawłowski, W., Szmeja, P., and

Wasielewska, K. (2017a). Semantic interoperability in

the Internet of Things: An overview from the INTER-

IoT perspective. Journal of Network and Computer

Applications, 81:111–124.

Ganzha, M., Paprzycki, M., Pawłowski, W., Szmeja, P., and

Wasielewska, K. (2017b). Streaming semantic transla-

tions. In System Theory, Control and Computing (IC-

STCC), 2017 21st International Conference on, pages

1–8. IEEE.

Ganzha, M., Paprzycki, M., Pawłowski, W., Szmeja, P.,

and Wasielewska, K. (2017c). Alignment-based

semantic translation of geospatial data. In 2017

3rd International Conference on Advances in Com-

puting,Communication Automation (ICACCA) (Fall),

pages 1–8.

Gyrard, A., Bonnet, C., and Boudaoud, K. (2014). Domain

knowledge interoperability to build the semantic web

of things. In W3C Workshop on the Web of Things,

pages 1–5.

Gyrard, A., Datta, S. K., and Bonnet, C. (2018). A sur-

vey and analysis of ontology-based software tools for

semantic interoperability in IoT and WoT landscapes.

In Internet of Things (WF-IoT), 2018 IEEE 4th World

Forum on, pages 86–91. IEEE.

Gyrard, A., Datta, S. K., Bonnet, C., and Boudaoud, K.

(2015a). Cross-domain Internet of Things application

development: M3 framework and evaluation. In Fu-

ture Internet of Things and Cloud (FiCloud), 2015 3rd

International Conference on, pages 9–16. IEEE.

Gyrard, A. and Serrano, M. (2015). A uniﬁed semantic en-

gine for Internet of Things and Smart Cities: From

sensor data to end-users applications. In 2015 IEEE

International Conference on Data Science and Data

Intensive Systems (DSDIS), pages 718–725. IEEE.

Gyrard, A. and Serrano, M. (2016). Connected Smart

Cities: Interoperability with SEG 3.0 for the Internet

of Things. In 2016 30th International Conference on

Advanced Information Networking and Applications

Workshops (WAINA), pages 796–802.

Gyrard, A., Serrano, M., and Atemezing, G. A. (2015b).

Semantic Web methodologies, best practices and on-

tology engineering applied to Internet of Things. In

Internet of Things (WF-IoT), 2015 IEEE 2nd World

Forum on, pages 412–417. IEEE.

Gyrard, A., Zimmermann, A., and Sheth, A. (2018). Build-

ing IoT-Based Applications for Smart Cities: How

Can Ontology Catalogs Help? IEEE Internet of

Things Journal, 5(5):3978–3990.

Howell, S., Rezgui, Y., and Beach, T. (2018). Water

utility decision support through the Semantic Web

of Things. Environmental Modelling & Software,

102(C):94–114.

Kitchenham, B. (2004). Procedures for performing sys-

tematic reviews. Keele, UK, Keele University,

33(2004):1–26.

Kitchenham, B. and Charters, S. (2007). Guidelines for per-

forming systematic literature reviews in software en-

gineering.

Kitchenham, B. A., Budgen, D., and Brereton, O. P. (2010).

The value of mapping studies: A participant-observer

case study. In Proceedings of the 14th International

Conference on Evaluation and Assessment in Software

Engineering, EASE’10, pages 25–33, Swindon, UK.

BCS Learning & Development Ltd.

Lohmann, S., Link, V., Marbach, E., and Negru, S. (2014).

WebVOWL: Web-based visualization of ontologies.

In International Conference on Knowledge Engineer-

ing and Knowledge Management, pages 154–158.

Springer.

Nagib, A. M. and Hamza, H. S. (2016). SIGHTED: a frame-

work for semantic integration of heterogeneous sensor

data on the Internet of Things. Procedia Computer

Science, 83:529–536.

Noy, N. F., McGuinness, D. L., et al. (2001). Ontology

development 101: A guide to creating your ﬁrst ontol-

ogy.

Pai, M., McCulloch, M., Gorman, J. D., Pai, N., Enano-

ria, W., Kennedy, G., Tharyan, P., and Colford, J. J.

(2004). Systematic reviews and meta-analyses: an il-

lustrated, step-by-step guide. The National medical

journal of India, 17(2):86–95.

Poveda-Villalón, M., Castro, R. G., and Gómez-Pérez, A.

(2014a). Building an ontology catalogue for smart

cities. In eWork and eBusiness in Architecture, Engi-

neering and Construction: ECPPM 2014, volume 1,

pages 829–839, Leiden. CRC Press.

Poveda-Villalón, M., Gómez-Pérez, A., and Suárez-

Figueroa, M. C. (2014b). OOPS! (OntOlogy Pitfall

Scanner!): An On-line Tool for Ontology Evaluation.

IoT Semantic Interoperability: A Systematic Mapping Study

543

International Journal on Semantic Web and Informa-

tion Systems (IJSWIS), 10(2):7–34.

Rhayem, A., Mhiri, M. B. A., and Gargouri, F. (2017).

HealthIoT Ontology for Data Semantic Representa-

tion and Interpretation Obtained from Medical Con-

nected Objects. In Computer Systems and Applica-

tions (AICCSA), 2017 IEEE/ACS 14th International

Conference on, pages 1470–1477. IEEE.

Serrano, M., Barnaghi, P., Carrez, F., Cousin, P., Verme-

san, O., and Friess, P. (2015). Internet of Things IoT

semantic interoperability: Research challenges, best

practices, recommendations and next steps. IERC:

European Research Cluster on the Internet of Things,

Tech. Rep.

Simperl, E. (2009). Reusing ontologies on the Semantic

Web: A feasibility study. Data & Knowledge Engi-

neering, 68(10):905–925.

Strassner, J. and Diab, W. W. (2016). A semantic interoper-

ability architecture for Internet of Things data sharing

and computing. In Internet of Things (WF-IoT), 2016

IEEE 3rd World Forum on, pages 609–614. IEEE.

Suárez-Figueroa, M. C. (2010). NeON Methodology for

Building Ontology Networks: Speciﬁcation, Schedul-

ing and Reuse. PhD thesis, Informatica. Ontology

Engineering Group.

W3C (2018). Semantic Integration & In-

teroperability Using RDF and OWL.

www.w3.org/2001/sw/BestPractices/OEP/SemInt.

Accessed: 20/02/2019.

Wohlin, C. (2014). Guidelines for snowballing in system-

atic literature studies and a replication in software en-

gineering. In Proceedings of the 18th International

Conference on Evaluation and Assessment in Soft-

ware Engineering, EASE ’14, pages 38:1–38:10, New

York, NY, USA. ACM.

Wohlin, C., Runeson, P., Höst, M., Ohlsson, M. C., Reg-

nell, B., and Wesslén, A. (2012). Experimentation in

software engineering. Springer Science & Business

Media.

ICEIS 2019 - 21st International Conference on Enterprise Information Systems

544