Future Generation Computer Systems 142 (2023) 1–3

Contents lists available at ScienceDirect

Future Generation Computer Systems

journal homepage: www.elsevier.com/locate/fgcs

Editorial

Semantic modeling and design patterns for IoT ecosystems

∗
Domenico Potena a , , Antonella Longo b , Alex Mircoli a , Marco Zappatore b
a
Dipartimento di Ingegneria dell’Informazione, Università Politecnica delle Marche, via Brecce Bianche, Ancona, 60131, Italy
b
Department of Engineering for Innovation, University of Salento, via Monteroni, Lecce, 73100, Italy

article info a b s t r a c t

Article history: In recent years, there has been an increasing diffusion of IoT-based devices that collect and share
Available online 22 December 2022 sensor data for a wide variety of applications. These data are highly heterogeneous as they are
Keywords:
obtained from various data sources utilizing different representation schemes, which represents a great
Internet of Things (IoT) richness in terms of stored information but also a strong limitation in terms of interoperability. Since
Industrial IoT traditional software modeling is not able to cope with this issue, several semantic approaches based
Semantic modeling on ontologies and/or linked data have been proposed in recent literature. Semantic modeling provides
Ontology a potential basis for interoperating among different systems and applications in the IoT. The semantic
Interoperability interoperability can be implemented for just one specific IoT device or for entire IoT ecosystems in
which different IoT devices interact with each other. The goal of this Special Issue is to provide insights
on recent advances in semantic modeling for IoT by presenting original scientific contributions in the
form of theoretical foundations, models, experimental research, and case studies.
© 2022 Published by Elsevier B.V.

1. Introduction designers who have to cope with the knowledge representation

stage in the IoT domain as well as to IoT system implementers
Internet of Things (IoT) applications are nowadays adopted who need to use and coordinate different IoT frameworks.
worldwide and they are expected to play an even more significant
role in the next future as technological enablers in several sectors, 2. Contents of the Special Issue
ranging from healthcare to smart working, from smart cities to
edge computing, thus creating new IoT ecosystems capable of All submitted manuscripts to this SI have undergone rigorous
supporting the novel use cases that are devised and deployed blinded review process and have been evaluated by a Committee
with a constantly growing frequency. of internationally recognized experts in this field (see Section 2.4)
Such an ongoing IoT-driven transformation demands for in- on the basis of coherence with the SI’s topics, originality and nov-
teroperating services, reliable sensor data integration, adequate elty, scientific, technical and technological soundness, method-
security levels, scalable access to resources and effective service ological robustness, state-of-the-art adequacy, consistency and
discovery methods. Therefore, when an environment as dynamic usefulness of any proposed use case, impact on the referenced
as the IoT is involved, knowledge representation and manage- scientific community and overall readability. A total of 10 articles
ment strategies face several challenges, in terms of capabilities
has been accepted, with an acceptance rate of 41.67%.
and requirements (e.g., data annotation, processing time, query-
The accepted articles showcased semantic and ontology-based
response time, assessment of missing values) to be guaranteed.
solutions addressing the IoT landscape in multi-faceted ways:
This special issue (SI) brings together original and high-quality
several contributions dealt with semantic models and frame-
research outcomes from academia and industry, all aimed at
works both for industrial and non-industrial IoT contexts. Another
advancing the current state of the art in this sector with novel so-
investigated perspective was the semantic interoperability in IoT
lutions either proposing semantic models for the (industrial) IoT
aimed at enabling effective data processing and exchange, while
domain or introducing novel IoT semantic design patterns, so to
the remaining papers presented other semantic approaches for
improve efficiency and reliability of IoT ecosystems. The articles
IoT.
in this SI, therefore, can help providing guidance to IoT system

∗ Corresponding author. 2.1. Semantic models and frameworks for Industrial IoT
E-mail addresses: [email protected] (D. Potena),
[email protected] (A. Longo), [email protected] (A. Mircoli), In ‘‘AoI-Aware Energy Control and Computation Offloading
[email protected] (M. Zappatore). for Industrial IoT’’, by J. Huang, H. Gao, S. Wan, and Y. Chen [1],

https://doi.org/10.1016/j.future.2022.12.025
0167-739X/© 2022 Published by Elsevier B.V.
D. Potena, A. Longo, A. Mircoli et al. Future Generation Computer Systems 142 (2023) 1–3

the authors focus on Industrial IoT (IIoT) by proposing an opti- framework for semantic annotation in IoT ecosystems. In partic-
mization method for the Age of Information (AoI) metrics, be- ular, they consider semantic annotation as a multi-label classifi-
cause of the significance of this metrics that introduces data cation problem, and develop an automatic annotation framework
semantics into traditional performance evaluation approaches. In based on multi-head self-attention mechanism. Experiments on
order to cope with typical IIoT challenges (e.g., limited compu- two large real-world datasets show high performance in semantic
tational resources and significant energy consumption), the au- annotation for mobile and edge services.
thors introduce a Deep Reinforcement Learning-based algorithm The article ‘‘A collaborative semantic framework based on
to optimize, as a dynamic optimization problem, the computa- activities for the development of applications in Smart Home
tion offloading and the energy control for both IoT devices and living labs’’, authored by C. Rodrìguez-Gallego, F. Dìez-Munoz,
edge servers, whose dynamic behavior is formulated via Markov M. Martìn-Ruiz, A. Gabaldòn, M. Dolòn-Poza, and I. Pau [6],
queueing models. To validate the model, a real-life telecommu- presents a collaborative framework for data extraction, charac-
nication dataset consisting of 7.2M Internet access trace data terization, and formalization. The goal is to enable the inter-
gathered from more than 3k base stations and referring to nearly pretation and reutilization of previously collected data related
10k mobile users is exploited, so to enable the comparison be- to Smart Home environments. The framework offers a set of
tween the proposed approach and a set of baseline alternative services, applications and mechanisms for proper synchronization
models. between elements and management of both events and activi-
The article ‘‘Process-aware IIoT Knowledge Graph: a semantic ties. Authors evaluated the framework on two projects and they
model for Industry IoT integration and analytics’’, authored by found that it gave significant benefits in terms of reusability and
C. Diamantini, A. Mircoli, D. Potena, and E. Storti [2], addresses interoperability.
the important role that semantic models can play in the Industry
4.0 context when the integration of significant data streams from 2.3. Semantic interoperability for data processing and data exchange
IIoT sources is required at cloud level and when data analyt- in IoT
ics are necessary. Starting from the ontological description of
IIoT sensors, processes, and KPIs, the authors propose a process- The article ‘‘Semantic models for IoT sensing to infer
aware knowledge graph aimed at enriching raw sensor data with environment–wellness relationships’’ by M. Zappatore, A. Longo,
information coming from two complementary perspectives: the A. Martella, B. Di Martino, A. Esposito, and S. A. Gracco [7],
one describing process activities and the one associated with the addresses the challenges in data collection, usage, and exchange
physical production environment where the IIoT sensors are de- originating from the heterogeneity and large-scale diffusion
ployed. The proposed conceptual knowledge layer is incorporated of non-standardized IoT solutions for personal and environmen-
in a framework allowing to query the knowledge graph and it is tal sensing, which reach the market with a constantly growing
instantiated in a case study dealing with the industrial production speed without being effectively supported by appropriate se-
of metal components. mantic interoperability capabilities. The authors propose a set of
In the article ‘‘Blockchain-based Trust Mechanism for Digital ontologies suitable to describe sensor-agnostic APIs for personal
Twin Empowered Industrial Internet of Things’’ by A. Sasiku- wellness monitoring as well as to model environmental monitor-
mar, V. Subramaniyaswamy, K. Ketan, V. Indragandhi, R. Lo- ing campaigns performed according to the mobile crowd sensing
gesh, G. Selvachandran, and A. Ajith [3], the authors combine a approach. The crucial correlation between these two perspectives
blockchain-based distributed network with a digital twin for In- is also considered and supported by semantic modeling patterns,
dustrial IoT (IIoT) applications. In particular, the article proposes so that complex functionalities based on inference rules can be
a blockchain-based Proof of Authority (PoA) trust mechanism to designed.
enhance security and data privacy. The approach is evaluated by In ‘‘A scalable, secure, and semantically interoperable client
simulating a blockchain network with a digital twin using IIoT for cloud-enabled Demand Response’’, by A. Cimmino, J. Cano-
sensor nodes. The simulation result shows that the proposed PoA Benito, A. Fernández-Izquierdo, C. Patsonakis, A. C. Tsolakis,
trust mechanism can reduce energy consumption and enhance R. García-Castro, D. Ioannidis, and D. Tzovaras [8], the need
security in IIoT. for standardized data models in Demand-Response (DR) sys-
In ‘‘Workload categorization for hazardous industries: the tems of services and IoT devices used by grid operators to mit-
semantic modeling of multi-modal physiological data’’, by J. L. igate changes in the energy supply, is considered. The authors,
König, A. Hinze, and J. Bowen [4], the authors address the peculiar tackle the heterogeneity in data formats and models of cur-
field of physical and cognitive fatigue for workers in hazardous rent and mainly centralized DR systems, which hamper data
industries, by proposing a semantic model for workload catego- exchange and limit system applicability and reliability on a larger
rization, so that fatigue-mitigation interventions can be identified scale, by proposing a distributed, cloud-based, semantic-enabled
and the risk of incidents lowered. The model, fed by multi- tool, named CIM. The tool allows distributing DR systems and
modal physiological parameters acquired via IoT and wearable components in the cloud, with a specific focus on security and
sensors monitoring the worker’s physical and health conditions, privacy for data exchange. The semantic-interoperability layer
classifies workloads into resting, cognitive, and physical thanks allows these decentralized systems to exchange data indepen-
to the integration of complex-event processing and rule based- dently from the adopted DR standard, thanks to normalized data
modeling. As a validation test case, several workers of different formulations and data payload translation mechanisms.
age groups are involved from the forest industry in New Zealand,
because of its prolonged workloads. The proposed categorization 2.4. Other semantic approaches for IoT
approach can be also considered as a suitable enabler for effective
workload prediction. The article ‘‘A fuzzy logic technique for virtual sensor net-
works’’, proposed by L. Caroprese, C. Comito, D. Talia, and E.
2.2. Semantic models and frameworks for non-Industrial IoT Zumpano [9], presents an approach to model virtual sensor net-
works based on fuzzy logic. The virtual sensor network can learn
In ‘‘Towards effective semantic annotation for mobile and the relationships between the physical quantities detected by the
edge services for Internet-of-Things ecosystems’’ by Y. Xu, W. sensors and model them through fuzzy rules, which provide an
Xiao, X. Yang, R. Li, Y. Yin, and Z. Jiang [5], the authors propose a approximation of the real values that can be used in case of
2
D. Potena, A. Longo, A. Mircoli et al. Future Generation Computer Systems 142 (2023) 1–3

sensor malfunction. The effectiveness of the proposed method- • Mahda Noura (Technische Universitat Chemnitz, Germany)
ology is also assessed by means of a thorough experimental • Majdi Rawashdeh (Princess Sumaya University for Technology,
evaluation. Jordan)
In ‘‘Semantics-based privacy by design for Internet of Things • Sanju Mishra (Universidad Autonoma de Tamaulipas, Mexico)
applications’’, by L. Alkhariji, S. De, O. Ran, and C. Perera [10], • Sergio Ilarri (Universidad de Zaragoza, Spain)
the authors focuses on the development of a personal assis- • Shahoua Wan (Shenzhen Institute for Advanced Study, China)
tant that can answer software engineers’ questions about Privacy • Viktorya Degeler (Technical University of Delft, The Nether-
by Design (PbD) practices during the design phase of IoT sys- lands)
tem development. This is achieved through the development of • Yudith Cardinale (Universidad Simon Bolivar, Venezuela)
the PARROT ontology, developed through a set of representative • Chiara Rucco (Università del Salento, Italy)
IoT use cases relevant for software developers. The ontology is
assessed within a user study that identified that the PARROT Declaration of competing interest
ontology can answer up to 58% of privacy-related questions from
software engineers. The authors declare that they have no known competing finan-
cial interests or personal relationships that could have appeared
Acknowledgments to the Committee of Reviewers to influence the work reported in this paper.

In order to achieve high-quality publication standards, a Com- References

mittee of Reviewers was specifically defined for this SI. Every
Committee members has been required to evaluate 2 papers at [1] J. Huang, H. Gao, S. Wan, Y. Chen, AoI-aware energy control and computa-
most, depending on the number of submitted papers. tion offloading for industrial IoT, Future Gener. Comput. Syst. 139 (2023)
Reviewers’ contributions, always timely and detailed, have 29–37.
[2] C. Diamantini, A. Mircoli, D. Potena, E. Storti, Process-aware IIoT knowledge
been pivotal for the outcome of the SI. We would like to acknowl- graph: A semantic model for industrial IoT integration and analytics, Future
edge the considerable time and efforts devoted by Committee Gener. Comput. Syst. 139 (2023) 224–238.
members to review submitted manuscripts, thus improving their [3] A. Sasikumar, S. Vairavasundaram, K. Kotecha, V. Indragandhi, L. Ravi, G.
quality significantly. We like to point out that many authors Selvachandran, A. Abraham, Blockchain-based trust mechanism for digital
twin empowered industrial internet of things, Future Gener. Comput. Syst.
of the reviewed manuscripts also have expressed their sincere
141 (2023) 16–27.
appreciation for the helpful and rigorous comments provided [4] J.L. König, A. Hinze, J. Bowen, Workload categorization for hazardous
by reviewers. It is a true pleasure, to thank the Committee of industries: The semantic modelling of multi-modal physiological data,
Reviewers: Future Gener. Comput. Syst. 141 (2023) 369–381.
[5] Y. Xu, W. Xiao, X. Yang, R. Li, Y. Yin, Z. Jiang, Towards effective semantic
• Adnan Anwar (Deakin University, Australia) annotation for mobile and edge services for internet-of-things ecosystems,
• Andrea Cimmino (Universidad Politécnica de Madrid, Spain) Future Gener. Comput. Syst. 139 (2023) 64–73.
[6] C. Rodríguez-Gallego, F. Díez-Muñoz, M.-L. Martín-Ruiz, A.-M. Gabaldón, M.
• Angelo Martella (Università del Salento, Italy) Dolón-Poza, I. Pau, A collaborative semantic framework based on activities
• Antonino Galletta (Università di Messina, Italy) for the development of applications in smart home living labs, Future
• Claudia Diamantini (Università Politecnica delle Marche, Italy) Gener. Comput. Syst. 140 (2023) 450–465.
• Davide Taibi (CNR, Italy) [7] M. Zappatore, A. Longo, A. Martella, B.D. Martino, A. Esposito, S.A.
Gracco, Semantic models for IoT sensing to infer environment–wellness
• Domenico Ursino (Università Politecnica delle Marche, Italy) relationships, Future Gener. Comput. Syst. 140 (2023) 1–17.
• Ellis Solaiman (Newcastle University, UK ) [8] A. Cimmino, J. Cano-Benito, A. Fernández-Izquierdo, C. Patsonakis, A.C.
• Emanuele Storti (Università Politecnica delle Marche, Italy) Tsolakis, R. García-Castro, D. Ioannidis, D. Tzovaras, A scalable, secure,
• Francesco Taglino (CNR, Italy) and semantically interoperable client for cloud-enabled demand response,
Future Gener. Comput. Syst. 141 (2023) 54–66.
• Grzegorz Nalepa (Jagiellonian University, Kraków, Poland) [9] L. Caroprese, C. Comito, D. Talia, E. Zumpano, A fuzzy logic technique
• Honghao Gao (Shanghai University, China) for virtual sensor networks, Future Gener. Comput. Syst. 137 (2022)
• Laura Genga (Eindhoven University of Technology, The Nether- 302–322.
lands) [10] L. Alkhariji, S. De, O. Rana, C. Perera, Semantics-based privacy by design
for internet of things applications, Future Gener. Comput. Syst. 138 (2023)
• Luca Virgili (Università Politecnica delle Marche, Italy) 280–295.