Interconnecting Urban Networks: A Novel Approach to Digital Twins

Through GlassBox Adaptation

Andrea Roberta Costagliola

1 a

, Marco Montanari

1 b

and Paolo Bellavista

2 c

Department of Computer Science and Engineering, University of Bologna, Italy

Alma Mater Studiorum, University of Bologna, Italy

Keywords:

Smart Cities, Urban Digital Twins, GlassBox, Hierarchical Models, Interoperability, Integration, Systems of

Systems, Multilayered Networks.

Abstract:

Smart cities are transforming urban living by leveraging technology and data to optimize urban services, to

improve resource efﬁciency, and to promote sustainability. Urban Digital Twins (UDTs) promise to play a

central enabling role in this transformation. However, state-of-the-art digital twin models still have to address

signiﬁcant challenging issues, particularly in terms of interoperability and integration with complex and mul-

tilayered (often legacy) urban systems. Some emerging approaches rely on speciﬁc standards and ontologies,

creating ”information silos” peculiar to each digital twin solution. After discussing the related and still open

technical challenges, this paper proposes a novel extension of the GlassBox urban simulation model to con-

ceptualize a city as a multilayered network, where nodes can be entities at different levels of granularity, such

as a single building, a urban energy network, or even an aggregated urban area. Our proposed solution aims

to ensure interoperability between these layers by enabling seamless real-time data exchange and optimized

resource management. Furthermore, integration between layers such as energy, transportation, and water is es-

sential to ensure data synchronization and provide the basis for more advanced smart city services, e.g., energy

consumption/production prediction. To practically exemplify the advantages of the proposed approach, our

innovative model is also illustrated when supporting a case study that focuses on urban transportation systems.

1 INTRODUCTION

Smart cities are revolutionizing urban living by lever-

aging technology and data to optimize services, to en-

hance resource efﬁciency, and to promote sustainabil-

ity. A cornerstone of this transformation is the devel-

opment of UDTs, i.e., dynamic digital replicas that

model real-world urban systems by using simulation

models coupled with data-driven models based on

machine learning fed by signiﬁcant ﬂows of real-time

data (Thelen et al., 2022). These tools enable simu-

lation, analysis, and prediction, by providing essen-

tial insights for planning and decision-making. Var-

ious global cities are developing UDTs for speciﬁc

needs: Singapore uses geospatial data and IoT for

urban planning and disaster management; Rotterdam

focuses on climate resilience with hydrological mod-

els; Cambridge optimizes transportation by analyz-

https://orcid.org/0009-0009-0119-2597

https://orcid.org/0000-0001-5026-6083

https://orcid.org/0000-0003-0992-7948

ing trafﬁc and mobility. In Italy, the Digital Twin of

Bologna stands out as a key initiative. Despite their

potential, current digital twin models face signiﬁcant

challenges, particularly in achieving seamless inter-

operability and in integrating complex multilayered

urban systems. Most approaches rely on speciﬁc stan-

dards and descriptors, leading to fragmented ”infor-

mation silos” that obstruct communication and data

sharing across systems. Addressing these issues re-

quires a ”system of systems” approach to fully cap-

ture the interdependencies of urban environments and

sub-systems, from individual buildings to city-wide

networks (Mihai et al., 2022).

In this context, simulation models emerge as a

foundational tool, offering simpliﬁed representations

of urban dynamics that can be scaled and enriched

with real-world data. Looking beyond traditional ur-

ban studies, insights can also be drawn from other

ﬁelds where simulation models have been success-

fully developed to represent complex systems. A no-

table example is the GlassBox model, whose develop-

ment started in 2011 by Maxis for its ﬂagship game

Costagliola, A. R., Montanari, M. and Bellavista, P.

Interconnecting Urban Networks: A Novel Approach to Digital Twins Through GlassBox Adaptation.

DOI: 10.5220/0013495100003953

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 14th International Conference on Smart Cities and Green ICT Systems (SMARTGREENS 2025), pages 193-204

ISBN: 978-989-758-751-1; ISSN: 2184-4968

193

SimCity, now in its 5th iteration (Maxis, 2012).

Urban simulation models like GlassBox, provide

a foundation for modeling urban dynamics through

simpliﬁed representations of units, resources, net-

works, and agents. However, GlassBox was designed

for simulation environments and struggles with the

variability and incompleteness of real-world data col-

lected in real deployment environments of smart

cities. To overcome these limitations, we propose an

original extension of the GlassBox framework, con-

ceptualizing cities as multilayered networks: nodes

within these networks represent various levels of ab-

straction/granularity, from individual buildings to ag-

gregated regions, and interact dynamically across lay-

ers such as energy, transportation, and water. In our

design/implementation work, a critical challenge we

had to address is enabling seamless communication

between these layers, allowing sub-systems with dif-

ferent protocols and standards to exchange data effec-

tively, which is crucial for optimized resource man-

agement and informed decision-making. Addition-

ally, we tackled the challenge of multilayer integra-

tion, by ensuring that interactions across layers—such

as energy demands inﬂuencing transportation or wa-

ter systems—are synchronized to achieve optimized

and coordinated responses.

For instance, the transportation network, encom-

passing both its physical (e.g., streets, trafﬁc lights)

and logical (e.g., public transport routes, schedules)

components, serves as a foundational layer in UDTs.

It integrates key urban services, such as electric ve-

hicles (EV), charging stations, and environmental in-

frastructures, enabling a holistic digital twin represen-

tation. By modeling these interactions, UDTs pro-

vide a high-level perspective of urban interdependen-

cies while ensuring the appropriate level of abstrac-

tion. This approach highlights the need to address in-

tegration challenges across multiple dimensions: hor-

izontal interoperability within the same network, ver-

tical multilayer integration across urban systems, and

organizational interoperability between governmental

structures.

In short and in summary, the primary contribu-

tions of this paper include:

• Analysis of UDTs from an interoperability and

multilayer integration perspective, addressing the

horizontal challenge of seamless data exchange

across diverse subsystems with different proto-

cols and the vertical challenge of synchronizing

interactions between urban layers to optimize re-

source management and reﬂect system interde-

pendencies.

• Extension of the GlassBox simulation framework

for urban environments, adapting it to model cities

as multilayered networks with varying levels of

abstraction, in order to overcome the limitations

of traditional simulation models and better repre-

sent the complexity of real-world urban systems.

• Proposal of a multilayered architecture that lever-

ages adaptable data description systems and struc-

tured data publishing endpoints. This architec-

ture is designed to handle the complexity of ur-

ban data ﬂows, facilitating the integration of both

simulation-related and real-time monitoring data

within the context of UDTs.

The remainder of the paper is organized as fol-

lows. Section 2 introduces the background and key

concepts of UDTs, with a rapid overview of the dig-

ital twin project of the city of Bologna, to contextu-

alize the challenges of interoperability and multilayer

integration. This section also reviews the state of the

art, by focusing on how existing solutions approach

these issues and outlining the foundational concepts

behind the GlassBox model. Section 4 explains how

the GlassBox model is originally extended in this pa-

per to address the speciﬁc needs of real-world urban

systems and demonstrates its application to the use

case of transportation in Bologna. In addition, this

section proposes an architecture for managing com-

plex urban data, along with an implementation of the

extended model, aimed at integrating both simulation-

related aspects and real-time monitoring data. Section

5 provides a discussion of the proposed model, by

evaluating its advantages and limitations in the con-

text of UDTs. Related work (Section 3) and conclu-

sive remarks and future work (Section 6) end the pa-

per, by outlining a few directions that we are currently

investigating as future work.

2 BACKGROUND

This section discusses key aspects of UDTs, inter-

operability as the basis for seamless communication,

multilayered integration as a means of managing in-

teractions between city sub-systems, and the impor-

tance of the GlassBox model as the basis for build-

ing a smart city framework. Together, these elements

highlight the opportunities and obstacles to designing

efﬁcient and scalable digital twins for smart cities.

2.1 Urban Digital Twins

An Urban Digital Twin is a virtual representation of

the physical assets, processes, and systems of a city

or community, powered by real data and constantly

updated. However, there is no universal approach or

SMEN 2025 - Special Session on Smart City and Smart Energy Networks

194

one-size-ﬁts-all model for their implementation: in

Italy, for example, the diversity of cities generates

as many digital twin models, each adapted to terri-

torial and operational peculiarities. This fragmen-

tation causes a lack of interoperability between sys-

tems, amplifying the challenges related to multilayer

integration, i.e., managing complex interactions be-

tween different layers of urban systems, such as en-

ergy, transportation, and waste management. The lack

of interoperability and the complexities of multilayer

integration highlight the practical challenges faced by

cities in implementing UDTs (Ferr

e-Bigorra et al.,

2022).

Globally, several cities worldwide have developed

Urban Digital Twins tailored to their speciﬁc needs.

Singapore’s ”Virtual Singapore” (Authority, 2014) in-

tegrates geospatial data, real-time IoT inputs, and

simulations to support urban planning and disaster

management. Rotterdam’s UDT (for BOLD Cities,

2025) focuses on climate resilience, using hydrolog-

ical models and environmental data to enhance ﬂood

risk management. In Cambridge, a digital twin (Cam-

bridgeshire, 2025) optimizes transportation and mo-

bility by analyzing trafﬁc ﬂow, public transit, and

pedestrian movement to inform urban planning deci-

sions.

Following these international examples, Italy is

also making strides in the ﬁeld of UDTs, with the Dig-

ital Twin project of Bologna standing out as a signif-

icant initiative. This project focuses on developing a

comprehensive Data Platform, hosted on CINECA’s

cloud infrastructure and leveraging FBK’s Digital-

Hub, to harmonize data from diverse systems and en-

sure seamless interaction between urban layers such

as energy, transportation, and healthcare. The plat-

form supports the collection, correlation, integration,

visualization, and analysis of city data, empowering

stakeholders like the Municipality of Bologna and its

subsidiaries to efﬁciently process and analyze data

from various sources, including legacy systems, IT

platforms, and IoT solutions. Developing Bologna’s

UDT involves several signiﬁcant challenges, particu-

larly during the design and development phases. One

major challenge is building an ontology for the UDT,

a ”semantic model” to optimize data acquisition, stor-

age, and analysis. Urban infrastructure, often oper-

ates independently, each following speciﬁc rules and

technologies.

For instance, in the context of mobility, difﬁculties

arise when attempting to integrate data from various

transportation systems, such as public transit sched-

ules, shared mobility services (e.g., bike-sharing, car-

sharing), and real-time trafﬁc information from IoT

sensors. Each of these systems may use different data

formats, protocols, and standards, making it complex

to create a uniﬁed view of the city’s mobility network.

Moreover, the multilayered nature of urban systems

becomes evident when examining how an increase in

EV adoption impacts both mobility and energy de-

mand. A rise in EV usage could strain the electrical

grid during peak hours, potentially leading to local-

ized shortages in charging availability. This, in turn,

inﬂuences transportation patterns, as users may ad-

just their routes or schedules based on the accessi-

bility of charging stations. Adding to the complexity

of these interconnected systems, non-electric vehicles

contribute to air pollution, which can be monitored

through sensors. The resulting data, often generated

in diverse formats, requires harmonization to be effec-

tively integrated into the broader dataset representing

the city’s dynamics.

2.2 Most Open Challenges in

Interconnecting Urban Networks

2.2.1 Interoperability

Interoperability refers to the capability of two or

more networks, systems, devices, applications, or

components to exchange and readily use informa-

tion securely, effectively, and with little or no in-

convenience to the user (Brutti et al., 2018). UDTs

and smart cities face signiﬁcant interoperability chal-

lenges across multiple levels. These challenges stem

from the heterogeneous nature of urban systems, each

with its own technologies and data standards (Quek

et al., 2021). Existing interoperability approaches can

be categorized into syntactic, semantic, technologi-

cal, and organizational dimensions (Hatzivasilis et al.,

2018), each addressing different layers of integration

challenges within complex urban ecosystems.

Syntactic interoperability focuses on establishing

common data formats and communication protocols

to enable basic data exchange between systems (Ay-

din and Aydin, 2020). Standardized data models,

such as XML or JSON, and widely used commu-

nication protocols like Message Queuing Telemetry

Transport (MQTT) protocol or REST APIs (Repre-

sentational State Transfer Application Programming

Interface) facilitate this layer of interoperability by

ensuring that data from one system can be transmitted

and understood by another. However, while syntac-

tic approaches address structural compatibility, they

do not resolve deeper issues related to the meaning

or context of the data that are exchanged (Veltman,

2001).

Semantic interoperability, on the other hand, aims

to ensure that data exchanged between systems are

Interconnecting Urban Networks: A Novel Approach to Digital Twins Through GlassBox Adaptation

195

not only structurally compatible but also contextu-

ally meaningful (Pliatsios et al., 2023). This involves

the use of shared vocabularies, ontologies, and meta-

data standards to deﬁne the meaning of data. For

example, semantic frameworks like the Web Ontol-

ogy Language (OWL) or Next Generation Service In-

terface - Linked Data (NGSI-LD)

, which supports

linked data and contextual information, allow differ-

ent systems to align their understanding of key con-

cepts, such as energy consumption or occupancy rate,

ensuring that data can be interpreted and utilized con-

sistently across diverse systems. Despite the advan-

tages, (Rahman and Hussain, 2020) underline signiﬁ-

cant challenges, including managing semantic hetero-

geneity due to differences in ontologies and terminol-

ogy adopted by different vendors, as well as the ab-

sence of universal standards for data representation.

Technological interoperability focuses on the

seamless integration of systems at the infrastructure

and protocol levels, ensuring that devices, networks,

and applications can communicate and exchange data

efﬁciently. This dimension addresses compatibility

in terms of hardware interfaces, network conﬁgura-

tions, and data transmission protocols. It often in-

volves adopting common communication technolo-

gies such as Bluetooth, Wi-Fi, or LoRaWAN (Long

Range Wide Area Network), as well as widely ac-

cepted network standards such as Transmission Con-

trol Protocol/Internet Protocol (TCP/IP). Technical

interoperability is essential to ensure that diverse sys-

tems within urban ecosystems, such as those in smart

cities or digital twins, can interact without requiring

signiﬁcant customization or modiﬁcations.

Organizational interoperability extends beyond

technical alignment to address the coordination of

processes, policies, and governance structures be-

tween different stakeholders and systems (Hardi et al.,

2023). This involves establishing agreements on data-

sharing policies, access control mechanisms, and pri-

vacy standards to enable secure and compliant data

exchange. Frameworks like the General Data Protec-

tion Regulation (GDPR) in the European Union high-

light the importance of organizational alignment, en-

suring that interoperability respects ethical and legal

boundaries. Collaborative governance models, such

as public-private partnerships, can also play a crucial

role in aligning the objectives and operations of dis-

parate urban systems.

Hybrid approaches that combine these dimen-

sions, are often necessary to tackle the layered com-

NGSI-LD is a speciﬁcation developed by ETSI (Eu-

ropean Telecommunications Standards Institute) for data

management and exchange in the context of smart cities and

the Internet of Things (IoT).

plexity of urban systems. For instance, middleware

platforms can act as intermediaries, providing syntac-

tic translation and semantic alignment while enforc-

ing organizational policies. Similarly, emerging tech-

nologies like data spaces and digital platforms, such

as Gaia-X in Europe (Tardieu, 2022), aim to create

ecosystems where interoperability is inherently sup-

ported through shared infrastructure and standards.

As a side effect of these factors, the lack of estab-

lished industry standards and common practices spe-

ciﬁc to UDT applications highlights the critical na-

ture of interoperability in this ﬁeld. This challenge is

ampliﬁed by the need to integrate diverse urban sys-

tems, ranging from transportation to energy grids, into

a holistic and uniﬁed digital representation.

2.2.2 Multilayer Integration

Multilayer integration poses a signiﬁcant challenge in

the creation of UDTs, as it involves managing the

complex interactions between various layers of ur-

ban networks while preserving the unique dynamics

and critical details of each layer (Peldon et al., 2024).

Urban systems are inherently multilayered, encom-

passing sectors such as energy, transportation, water,

waste management, and more. Each of these layers

operates with distinct physical infrastructures, func-

tional dynamics, and temporal behaviors, yet they are

deeply interconnected (Aleta et al., 2017). The chal-

lenge lies in capturing these interdependencies in a

uniﬁed framework without oversimplifying individual

layers or losing the granularity necessary for accurate

analysis and decision-making.

An approach to multilayer integration is the adop-

tion of hierarchical modeling frameworks (Lu et al.,

2020), where each layer is represented with a level

of abstraction appropriate to its function, while main-

taining links to more detailed submodels. For exam-

ple, energy systems might be modeled at a high level

as grids or hubs, but with the ability to drill down into

ﬁner details, such as individual solar panel outputs or

battery storage levels, when needed. However, ag-

gregated models at higher levels may overlook im-

portant local dynamics, such as ﬂuctuations in solar

panel output or energy storage levels, which can lead

to inaccuracies in predictions.

Data synchronization and alignment are also criti-

cal for multilayer integration (Shih et al., 2015). Dif-

ferent layers often operate on diverse temporal and

spatial scales—transportation systems might gener-

ate data in seconds, while water management systems

might use hourly or daily data. Integrating such data

streams requires techniques like temporal resampling,

spatial aggregation, and interpolation to harmonize

datasets without losing essential information.

SMEN 2025 - Special Session on Smart City and Smart Energy Networks

196

Additionally, cross-layer optimization algorithms

play an important role in achieving effective integra-

tion. These algorithms enable the UDT to analyze

and manage trade-offs between layers, ensuring that

decisions in one domain consider their impact on oth-

ers (Castelli et al., 2019). For example, optimiz-

ing energy usage might involve not only balancing

supply and demand but also adjusting transportation

schedules to reduce peak loads on the grid. How-

ever, cross-layer optimization algorithms have to face

still open challenges such as computational complex-

ity and conﬂicting objectives between different do-

mains.

Finally, the visualization of multilayer interac-

tions is a key aspect of achieving a comprehensive

understanding. Advanced visualization tools, such as

3D city models or interactive dashboards, can help

stakeholders navigate the complexity of integrated

systems. These tools provide intuitive representations

of data ﬂows, dependencies, and scenarios, empow-

ering planners and decision-makers to explore the ef-

fects of interventions across layers. However, a sig-

niﬁcant limitation of many current visualization sys-

tems is their passive nature—they often display multi-

layer interactions but do not allow users to directly in-

teract with or manipulate the data. This lack of inter-

activity restricts the ability to test scenarios, explore

dynamic responses, or implement real-time interven-

tions, thereby reducing the practical utility of the in-

sights provided.

Further complexity is given by coordination

across neighbor operators that manage complex dig-

ital twin infrastructures with speciﬁc integration for-

mats. The lack of a de-facto industry standard, as pre-

viously noted, makes it very complex to design and

implement efﬁcient interactions across digital twins.

2.3 Glassbox

The Glassbox Simulation Engine is an advanced sim-

ulation engine developed by Maxis in 2011, for use

in SimCity (2013). Presented at GameDeveloperCon-

ference 2012 it displayed an innovative and general-

purpose approach to the simulation environment. De-

signed to provide a rich and dynamic simulation ex-

perience, Glassbox stands out for its ability to manage

large volumes of data and create highly reactive and

interconnected virtual systems in a 2013 technologi-

cal infrastructure. With its ﬂexible architecture, the

engine was conceived not only to support the iconic

city-building game but also a variety of other simula-

tion titles, making it a versatile tool for creating com-

plex interactive experiences.

The core of Glassbox operations is the detailed

management of resources, units, maps, networks,

agents, and rules (Figure 1). Resources, such as

oil, electricity, wood, and water, are fundamental el-

ements within the game, managed through containers

called ”bins” that track their quantity and distribution.

Each unit, representing entities such as houses or fac-

tories, interacts with resources through the rules that

describe the needs and conditions speciﬁc to each unit

in the game as well as its production. These needs and

resources move along speciﬁc networks. The sim-

ulation is further enriched by the presence of maps,

which represent environmental variables such as the

availability of natural resources, pollution, and land

desirability, in general factors that inﬂuence the deci-

sions of agents and the evolution of the virtual system.

Unit

Resource

Agent

Network

Map

Issued/Consumed

Emits/Consumes

Collects

Delivers

Moves

Accepts

It is

connected

It is linked

Traces

It is

monitored

Figure 1: Glassbox Model.

One of the distinctive features of Glassbox is its

ability to easily adapt to new game dynamics, thanks

to a system that manages behavior rules deﬁned by

customizable scripts. These rules allow for accu-

rate simulation of resource transfer and transforma-

tion, creating a complex and interconnected ecosys-

tem where every action has a tangible impact on the

game environment. This approach, which integrates

real-time data management with the simulation of

emergent behaviors, enables players to interact with

a virtual world that dynamically responds to their

choices.

3 RELATED WORK

The growing complexity of urban systems and the in-

creasing reliance on digital twins in smart city frame-

works have highlighted interoperability and multilay-

ered integration as critical challenges (Atkinson et al.,

2022). Existing solutions often fail to enable seam-

less integration and communication between differ-

ent platforms and models, limiting their potential to

create interconnected urban ecosystems (Quek et al.,

2021). Among the solutions proposed to address

Interconnecting Urban Networks: A Novel Approach to Digital Twins Through GlassBox Adaptation

197

this issue is FIWARE (FIWARE Foundation, 2024),

an open-source platform that leverages standardized

frameworks, such as NGSI-LD, to enable the manage-

ment and exchange of context data (Bauer, 2022). FI-

WARE aims to provide a ﬂexible and scalable founda-

tion for developing urban applications, yet its integra-

tion with other platforms often uncovers signiﬁcant

limitations. The integration of FIWARE with other

platforms, such as robotic operating systems like ROS

2, requires signiﬁcant development effort, especially

in transforming and mapping data between disparate

formats and standards. Additionally, while NGSI-LD

offers a structured approach to context data manage-

ment, its reliance on speciﬁc standards can result in

incompatibility with alternative systems or legacy in-

frastructures that employ different schemas or data

structures (Viola et al., 2019; Abid et al., 2022; Ku-

mar et al., 2022). This dependency can impede seam-

less data exchange, necessitating additional transfor-

mations that add to the complexity and maintenance

overhead of these systems. Moreover, maintaining

and updating the libraries required for these transfor-

mations introduces further challenges, as any evolu-

tion in data standards or platforms may demand cor-

responding updates, thereby increasing the workload

for developers and system operators.

UrbanSim (Waddell et al., 2018) and CityZenith

(Mukherjee et al., 2014) exemplify the interoper-

ability and multilayered integration challenges faced

by urban modeling platforms. UrbanSim excels in

simulating urban dynamics by integrating land use,

transportation, and economic data, but its reliance on

highly detailed, domain-speciﬁc datasets complicates

integration with other systems and increases compu-

tational demands, particularly for large metropolitan

areas. Similarly, CityZenith provides advanced 3D

visualization and analytical tools for city management

but remains limited in scalability and interconnectiv-

ity, restricting its effectiveness in enabling inter-city

collaboration and integrated urban planning. Ontolo-

gies play a crucial role in addressing interoperabil-

ity challenges by providing a structured framework

for representing and integrating diverse data sources.

They deﬁne a common vocabulary and relationships,

enabling different systems to exchange and interpret

information effectively. However, semantic interop-

erability often remains limited due to the fragmented

nature of urban systems, where data is organized in

domain-speciﬁc ways without a shared framework.

This creates barriers for cross-sector communication,

particularly in smart cities. Additionally, integrat-

ing multiple ontologies or aligning them across sys-

tems demands signiﬁcant resources, as each system’s

unique schemas require complex and time-consuming

adaptations(Karabulut et al., 2024). CityGML is

a standard for representing three-dimensional urban

models that aims to facilitate interoperability and

multilayered integration. However, practical imple-

mentation of CityGML often encounters difﬁculties

due to variations in the ontologies and data schemas

employed by different systems, which can lead to

inconsistencies in data exchange(Buyuksalih et al.,

2017). Ontological frameworks like RDF (Resource

Description Framework) offer a ﬂexible mechanism

for representing geospatial data, but they frequently

require complex customization to integrate real-time

data streams, such as those related to urban mobil-

ity or environmental monitoring.g. These adjustments

demand signiﬁcant expertise and computational re-

sources, further complicating efforts to establish in-

teroperable and multilayered networks of UDTs.

4 OUR ORIGINAL PROPOSAL

FOR EXTENDING THE

GLASSBOX MODEL

Our proposal originates from the Glassbox basic

model, but takes into consideration that, while Glass-

box was designed to act on a simulation environment

where players could never ”rewind” and see what a

different decision would have brought, a UDT has to

cover both simulation-related aspects as well as in-

the-ﬁeld ﬂows of monitoring data, which can not be

presumed to be in real time but often aggregated over

non homogeneous time periods.

For these reasons our Glassbox-extended model

uses the same concepts (resources, units, maps, net-

works, agent and rules) but adds the ”metrics” aspect

to them, deﬁning snapshots of the various elements

valid for speciﬁc timeframes. For example, a unit

would not just be an entity associated with a given

set of resources and a given set of networks, but also

deﬁne a metric, a value and a timeframe for which

that value is valid. The same happens with the maps:

the global situation is deﬁned by temporal and spa-

tial aggregations that represent the current state of the

system. Thus, maps are not limited to describing ge-

ographic locations or the physical distribution of re-

sources, but incorporate a set of temporal metrics as-

sociated with each node in the network. For example,

an energy map might include data on generation ca-

pacities, consumption ﬂows, and storage levels, each

with values valid for speciﬁc time intervals, such as

hours, days, or weeks.

Similarly, networks are no longer static, but dy-

namic, updated through real-time data streams or pe-

SMEN 2025 - Special Session on Smart City and Smart Energy Networks

198

riodic aggregates. Each link in a network is enriched

with parameters that deﬁne the context of interactions,

such as maximum transport capacity, energy losses,

or latency times. These parameters are constantly re-

assessed in light of systemic changes, allowing more

realistic simulations and prediction of bottlenecks or

inefﬁciencies.

Building upon this dynamic nature, our model

conceptualizes the network as not just a representa-

tion of connections and ﬂows, but as a modular frame-

work that can operate at multiple scales. A critical ad-

vantage of this approach is its ability to support cross-

city interoperability. Using the inherent ﬂexibility of

the network model, entire systems, such as energy

grids or transportation networks, can be abstracted

into a single node capable of emitting and receiving

resources. This abstraction enables seamless integra-

tion between urban systems of varying complexity

or granularity, fostering scalability and collaboration

across different cities or regions. At the same time,

the model retains the capacity to analyze and opti-

mize interactions within each subsystem, maintaining

a balance between global and local perspectives. To

support this dual-level functionality, the model incor-

porates global and local metrics as key elements for

capturing and analyzing system states. Global met-

rics provide a macroscopic view by summarizing the

overall state of the system at regular intervals, form-

ing the basis for historical analysis and validation of

simulation results. Local metrics, in contrast, detail

the speciﬁc states of subsystems or individual nodes,

such as buildings or electric vehicle charging stations,

enabling precise optimization and decision-making at

a granular level.

Finally, the concept of “rules” has been expanded

to integrate more complex conditional logics that gov-

ern interactions among system components. These

rules deﬁne not only the ﬂow of data between entities

but also incorporate temporal constraints and critical

thresholds, reﬂecting both urban realities and simu-

lation requirements. For instance, a rule might spec-

ify that during peaks in energy demand, certain urban

sectors should be prioritized for energy distribution

based on dynamically calculated criticality metrics.

4.1 Architecture

In order to deﬁne a ﬂexible architecture that can be

helpful in a complex data landscape like the UDT, it

is important to leverage both adaptable data descrip-

tion systems as well as structured data publishing end-

points. Figure 2 illustrates this architecture, which is

structured into four main layers:

• Presentation Layer:This layer is responsible for

interfacing with external systems and users, en-

suring that data are both ingested and accessed in

a standardized and meaningful way. It is divided

into two sub-layers:

– Semantic Ingestion Layer: this sub-layer fo-

cuses on the semantic description of datasets

and data streams entering the platform.

– Uniﬁed Data Access Layer: this sub-layer

provides standardized access to data for exter-

nal systems or users. It ensures that data output

is consistent, secure, and accessible via well-

deﬁned APIs or protocols.

The Semantic Ingestion Layer enriches incom-

ing data with metadata and semantic descriptions,

while the Uniﬁed Data Access Layer uses this in-

formation to provide consistent and structured ac-

cess via standard APIs and protocols.

• Application Layer: this layer is the core of data

processing, where ETL (Extraction, Transforma-

tion, and Load) operations are deﬁned and ex-

ecuted. In addition,there are modules for plan-

ning and coordinating workﬂows, making sure

that ETL processes are executed at the right time

and in the right order;

• Data Layer:this layer is responsible for manag-

ing, storing, and accessing raw and transformed

data. Its main function is to ensure that informa-

tion is organized in a scalable, secure and efﬁcient

manner, supporting analysis, simulation and visu-

alization operations.

• Integration Layer: this is a vertical cross-cutting

layer that spans across all other layers. It is de-

signed to facilitate advanced data operations, such

as simulations, evaluations, and integrations with

external systems. More speciﬁcally, it is respon-

sible for elaborating the rules to apply in simu-

lations and advanced data operations. The simu-

lation context depends on the descriptions of the

rules and semantic metadata from the Semantic

Ingestion Layer, on raw and processed data from

the Data Layer, and on transformation logic and

processing workﬂows from the Application Layer.

Once the simulation context is ready, the Integra-

tion Layer communicates with the Uniﬁed Data

Access Layer to ensure that the processed data

and results are made accessible to external sys-

tems or users in a standardized and secure way.

Moreover, this layer monitors and tracks the ﬂow

of data and the operations performed in the vari-

ous layers.

Interconnecting Urban Networks: A Novel Approach to Digital Twins Through GlassBox Adaptation

199

Extraction

Transformation

Scheduling &

Orchestration

Tracking &

Logging

Backup Systems

Data Lake

Data Layer

Application Layer

Presentation Layer

Integration Layer

Semantic Ingestion LayerUnified Data Access Layer

Data Warehouse

Elaborates

rules

Figure 2: Architecture of the extended Glassbox Model.

4.2 Implementation

Following the structure of the previously deﬁned ar-

chitecture, the core of the infrastructure is a Medi-

aWiki installation with a semantic module, which en-

ables the use of semantically enhanced templates for

creating data descriptors. This forms the foundation

of the Semantic Ingestion layer within the Presenta-

tion layer, providing tools for non-technical operators

to describe datasets in various formats. The templates

facilitate the correct deﬁnition of the many descrip-

tive aspects of the model, enabling both internal and

external descriptor deﬁnitions. In addition to the def-

inition of the models, the tool also collects the rules

for the simulation itself.

The Uniﬁed Data Access layer, also part of

the Presentation layer, ensures that data is exposed

through standard APIs and formats, making informa-

tion reusable and integrable into common libraries

and tools. Examples include the publication of maps

as Web Map Service (WMS) raster layers, allowing

navigation of complex datasets as simple visual lay-

ers, while maintaining ﬂexibility for reuse in new,

complex visualizations. Real-time calculations and

updates are made available via MQTT, supporting

both 2D and 3D visualizations of simulated items.

The Application layer is where workﬂows for

data ingestion, transformation, and distribution are

managed. Tools such as OpenMetadata and CKAN

(CKAN, 2025) (Comprehensive Knowledge Archive

Network), operating within this layer, enable seam-

less internal and external descriptor deﬁnitions.

OpenMetadata supports internal use cases, while

CKAN facilitates external data sharing. These tools

ensure that data ingestion and management processes

are streamlined, supporting a wide range of applica-

tions and user needs.

The Data layer is responsible for raw data storage.

MinIO, an Amazon S3 replacement, is used in this

layer to store temporary datasets. Geographic data is

stored for example in the netCDF format, enabling

efﬁcient management and retrieval of structured data,

while purely tabular data is stored in parquet format.

This layer provides the foundational infrastructure re-

quired for handling large volumes of data, supporting

both simulation and visualization processes.

The Integration layer handles backend calcula-

tions and the simulation engine, enabling complex

data evaluations. Libraries such as pandas, geopan-

das, and movingpandas help processing and analyz-

ing data, ensuring that disconnected simulations can

be visualized in various formats through the Uniﬁed

Data Access layer. The simulation itself is based on a

custom developed engine, relying on the original con-

cepts of GlassBox but enabling the tracing of the met-

rics. This is important for the quality analysis of the

model as well as of the collected data: a discrepancy

of the metrics could have origin in a wrong descrip-

tion of the dataset, but also in an incomplete cover-

age of the data available in the platform. This ap-

proach decouples the simulation engine from the vi-

sualization process, enabling ﬂexible visual represen-

tation, whether through 3D analysis tools or simpler

2D maps for broader accessibility.

Finally, the infrastructure also supports the dy-

namic growth of an emerging ontology for the city.

As new datasets are integrated and new requirements

arise, the ontology evolves to interconnect data over

time, progressively enriching the city’s description

and supporting an adaptable data ecosystem.

4.3 Model Applied to the Use Case

The use case we will analyze is mobility, which is

one of the core infrastructures of the city. This use

case focuses on how urban mobility systems interact

dynamically with other urban layers, such as energy

and environmental factors. Using the extended Glass-

Box model, it is possible to describe its elements with

precision. The units are all nodes that interact with

the networks. Examples include:

• Trafﬁc Spire: Sensors located along roads that

monitor the ﬂow of vehicles in real time. Each

Trafﬁc Spire is a unit that collects data on vehi-

cles and records them as resources (number of ve-

hicles).

• Charging Stations: Units that provide power to

EV. The associated resource is the electrical ca-

pacity available for charging.

Each of these units interacts with the networks

through deﬁned rules, contributing to data collection

and analysis. For example, trafﬁc spires measure ve-

hicle ﬂow and serve as input to air quality calcula-

SMEN 2025 - Special Session on Smart City and Smart Energy Networks

200

tions. Similarly, EV chargers interact with the elec-

tric grid to monitor energy consumption and demand

patterns.

Resources tracked include:

• Car (Car.carbon and Car.electric): Each car is

a numeric entity that is tracked through the road

network. The entry of a car into a unit, such as a

Trafﬁc Spire, is recorded as an incoming resource.

• Electricity: The capacity of charging stations and

of the electric network is limited.

Maps are visualizations that represent these re-

sources, providing a comprehensive view of the entire

urban system.

• Air Quality Map: Shows pollution levels in dif-

ferent areas of the city.

• Noise Map: Shows noise pollution levels in dif-

ferent areas of the city.

Networks in this context represent the set of con-

nections between units that collect, process, and share

information or resources within the city. The net-

works involved include

• Road Network: Representing the infrastructure

for vehicles, pedestrians, bicycles, and public

transport.

• Electric Grid: Supporting electric vehicle (EV)

infrastructure, including charging stations and en-

ergy distribution.

Municipality open or internal data sets are often

structured as CSV ﬁles in various formats, which can

be normalized according to the Extended GlassBox

Model, as described in Table 1.

Table 1: Deﬁnition of a Trafﬁc spire csv ﬁle - Unit.

Field Type Description

ID String Identiﬁer for the spire

Position GeoPoint Coordinates of the spire

Time Timestamp Timestamp of the

beginning or end

of the collected

metric

Vehicles Integer # of vehicles sensed by

the device in the last

valid timeslot

Accuracy Float % of validity of the

collected metric

(Vehicles)

Table 2 represents the csv ﬁle structure for the data

deﬁnition of a static sensor-based Air Quality Map.

In a typical smart city scenario, Trafﬁc Spires

along the urban roadways detect an increase in traf-

ﬁc.

Table 2: Deﬁnition of an air quality csv ﬁle - Map.

Field Type Description

Area GeoPolygon Area for the

data collection

Station GeoPoint Location of the

station

StationName String Name of the

station

Pollutant String Identiﬁer of

the measure

Time Timestamp Timestamp of the

beginning or end

of the collected

metric

Value Float # measure of the

pollutant during the

valid timeslot

Listing 1 shows a practical representation of of

a rule applied to the Trafﬁc Spire unit. As a car

(via) enters the area monitored by a Trafﬁc Spire,

it is immediately detected by the sensor (appliesTo).

The Trafﬁc Spire records the number of vehicles

passing through (log resource: Vehicles and Vehi-

cle.{type}), along with the timestamp for each pas-

sage (-timestamp=Time). This means that as the ve-

hicle moves through the monitored area, it is counted

as an entering resource (local Car.{type} in 1) and

contributes to the overall trafﬁc ﬂow data. Once the

vehicle enters the area, it is ”destroyed” in the sense

that its data is ﬁnalized and recorded as an exiting re-

source, which helps calculate the impact on the trafﬁc

conditions at that moment. A new vehicle is generated

and sent as output of the node itself (agent Car.{type}

out 1).

Listing 1: Unit Rule Code Example - Spire Trace.

unitRule s p i r e T r a ce

applies T o T r af f i c S p i r e

via C a r

using Road N e t w o r k

local Car .{ ty p e } in 1

age n t C ar .{ ty p e } ou t 1

lo g - timesta m p = T ime Vehicles 1

lo g - timesta m p = T ime Vehicles . { t y pe } 1

en d

At the same time, air quality monitoring stations

report a rise in PM2.5 levels, measuring safety limits.

Listing 2 represents the rule applied to a single agent

during its lifetime deﬁning that every 10 timeslots (re-

peatAfter 10) if the agent contains a carbon-fueled

car (condition resource=Car.carbon), the AirQuality

map(-map=AirQualty) will increase its measure of

pollutants by a speciﬁc amount in a speciﬁc area

round the position of the agent(area pm10 10 5, area

Interconnecting Urban Networks: A Novel Approach to Digital Twins Through GlassBox Adaptation

201

pm5 15 10, area pm25 12 15). This generated air qual-

ity map can be compared to the sensor based air qual-

ity map coming from institutional providers in order

to validate the simulation.

Listing 2: Unit Rule Code Example - Car Pollution.

agentRu l e c a rP o l l u t i o n

conditi o n r e s o u r c e = C ar . c a r b o n

rep e a t A f t e r 1 0

ar e a - ti m e s t a m p = T i me - map = A i rQua l i t y p m 10 10 5

ar e a - ti m e s t a m p = T i me - map = A i rQua l i t y p m5 15 10

ar e a - ti m e s t a m p = T i me - map = A i rQua l i t y p m 25 12 15

ar e a - ti m e s t a m p = T i me - map = N o i s e M a p noi s e 8 1 0

en d

Charging stations attract EV, which, in order to

refuel, generate increased trafﬁc in their vicinity, of-

ten critical during rush hour or in areas with limited

road infrastructure. This congestion involves not only

electric vehicles, but also internal combustion vehi-

cles, which, stuck in trafﬁc, release pollutants such as

PM2.5 and NOx, worsening air quality.

Moreover, during peak trafﬁc periods, charging

stations can experience very high energy demand, cre-

ating pressure on the remaining capacity of the lo-

cal electric grid, especially if multiple stations in the

same area are overloaded. Pollution data, resulting

from road congestion, can inform policies to encour-

age electric vehicle adoption, such as charging incen-

tives or the introduction of low-emission zones (LZs)

in highly polluted areas. An example of a unit rule

applicable to charging stations can be seen in Listing

Listing 3: Unit Rule Code Example - Charging Station.

unitRule c h a r g i n S t a t i o n

applies T o C ha r g i n g S ta t i o n

vi a Car

usi n g Tr a f f i c N e t w o r k

usi n g Ele c t r i c G r i d

loc a l C ar . el e c t r i c i n 1

loc a l Elec t r i c i t y 10 0 0

wa i t 1 4 4 00

age n t C ar . el e c t r i c ou t 1

lo g - timesta m p = T ime - event = in C ar . el e c t r i c 1

lo g - timesta m p = T ime - event = o ut Ca r . e l e c t r i c 1

en d

This process allows tracking of the interactions

within and across networks over time. Addition-

ally, the collected metrics support the validation of

the simulated model. This interconnectedness be-

tween networks and data structures facilitates com-

plex decision-making and validation processes.

5 DISCUSSION

The extended model we propose addresses both hor-

izontal and vertical challenges in an integrated man-

ner, focusing on solving interoperability and multi-

level integration issues within a complex urban dig-

ital context. To tackle the horizontal interoperabil-

ity challenge, our approach introduces the concept of

metrics for each entity (units, networks, maps) within

the urban network. The metrics are not only numer-

ical values but also integrate temporal and spatial di-

mensions, providing dynamic validity to the informa-

tion. For example, considering the previously dis-

cussed mobility example, the ﬂow of vehicles through

an intersection is represented not only as a number

(e.g., the number of cars per hour) but also with a tem-

poral connotation (e.g., during rush hour) and spatial

connotation (e.g., in a speciﬁc area). This approach

enriches the data with contextual information, making

it more meaningful and facilitating the integration of

mobility systems using heterogeneous protocols and

formats, such as sensor data, trafﬁc management APIs

and public transportation systems.

The simple GlassBox model acts as a “lingua

franca” that facilitates data exchange between tech-

nically diverse systems. The temporal and spatial

metrics added by our proposed extension help over-

come semantic barriers by contextualizing the data

and aligning systems with different ontologies or vo-

cabularies. In this way, semantic interoperability is

not just about ”sharing” data, but about understanding

and consistency in interpreting information, improv-

ing its usability and the ability to integrate it without

losing meaning.

Although metrics do not directly solve organiza-

tional challenges, they provide a solid technical foun-

dation for collaboration among different stakeholders,

such as public transportation authorities, private oper-

ators, and urban infrastructure managers. The stan-

dardization, both syntactic and semantic, enabled by

metrics reduces ambiguities and conﬂicts that typi-

cally hinder cooperation among different entities, pro-

moting the adoption of interoperable frameworks at

the organizational level as well. Additionally, the in-

troduction of data management rules, which include

the temporal and spatial validity of information, helps

apply clearer management policies, increasing the re-

liability of shared data and the level of collaboration

among various stakeholders.

Regarding the vertical challenge of multilayer in-

tegration, our model focuses on managing the inter-

actions among the different components of the urban

system without sacriﬁcing the necessary granularity

for accurate analysis. In this context, the idea of “net-

SMEN 2025 - Special Session on Smart City and Smart Energy Networks

202

works of networks” ﬁts as a key principle: each net-

work (such as energy, transportation, water) is not just

seen as an isolated entity, but as part of a larger net-

work that dynamically interacts with other networks.

Local metrics make it possible to analyze the behavior

of individual units (e.g., energy ﬂows at an EV charg-

ing station), while global metrics provide an overall

view of the state of the system, enabling forecasting

and resource optimization at the macro level. These

metrics, which are constantly updated, allow the dy-

namics between layers to be synchronized, overcom-

ing the challenges associated with data from systems

operating on different temporal and spatial scales. In

addition, inter-domain optimization rules, which bal-

ance the trade-offs between different domains, allow

us to address cross-domain optimization challenges

while maintaining a holistic view of urban interdepen-

dencies. In this way, our model concretely addresses

the difﬁculties arising from complex multilevel inter-

actions, enhancing the ability to make timely and in-

formed decisions.

6 CONCLUSIVE REMARKS AND

FUTURE WORK

In this work, we have proposed an extended digital

twin model based on an existing simulation model

called GlassBox, in order to address the challenges

of interoperability and multilevel integration in com-

plex urban contexts. Our approach relies on introduc-

ing the concept of metrics for each entity within the

urban network, while integrating temporal and spa-

tial dimensions. This enrichment allows for a more

precise and meaningful representation of information

from heterogeneous systems, such as public transport

or shared mobility systems. The proposed model aims

to overcome the traditional limitations of digital twins

by offering a holistic view of the city, which facili-

tates the integration of different urban infrastructures

and networks.

The model presents several advantages. First, be-

ing simple and based on bottom-up modeling, it gen-

erates limited additional work/overhead and allows

for scalable adoption in different urban contexts. The

ability to model complex informational structures lin-

early enables addressing urban situations with various

needs and characteristics, without compromising the

consistency of the system. Finally, one of the most in-

novative aspects of this model is its ability to promote

interoperability between UDTs, as each network can

be represented by a single point, simplifying commu-

nication between diversiﬁed sub-systems.

However, our original model also has some lim-

itations. The integration of data and metrics, while

crucial for a comprehensive view, adds complexity to

the model itself, making the management and main-

tenance of simulations more challenging, especially

when dealing with constantly evolving environments.

Looking ahead, the next step will be to conduct a

wide set of experiments to assess the effectiveness of

our original model in real-world contexts. Validating

the results obtained from simulations and case stud-

ies coming from cities, as well as the application of

the model to datasets regarding the past, will provide

valuable data to further reﬁne our model.

REFERENCES

Abid, A., Lee, J., Le Gall, F., and Song, J. (2022). To-

ward mapping an ngsi-ld context model on rdf graph

approaches: A comparison study. Sensors, 22(13).

Aleta, A., Meloni, S., and Moreno, Y. (2017). A multi-

layer perspective for the analysis of urban transporta-

tion systems. Scientiﬁc Reports, 7(1):44359.

Atkinson, R. A., Zaborowski, P., Noardo, F., and Simonis,

I. (2022). Smart cities – systems of systems inter-

operability and ogc enablers. ISPRS Annals of the

Photogrammetry, Remote Sensing and Spatial Infor-

mation Sciences, X-4/W3-2022:19–26.

Authority, S. L. (2014). Virtual singapore: A 3d city model

platform for knowledge sharing and community col-

laboration. Accessed: 2025-02-21.

Aydin, S. and Aydin, M. N. (2020). Semantic and syntactic

interoperability for agricultural open-data platforms in

the context of iot using crop-speciﬁc trait ontologies.

Applied Sciences, 10(13).

Bauer, M. (2022). Fiware: Standard-based open source

components for cross-domain iot platforms. In 2022

IEEE 8th World Forum on Internet of Things (WF-

IoT), pages 1–6.

Brutti, A., Frascella, A., Gessa, N., de sabbata, P., and Nov-

elli, C. (2018). Interoperability in the smart city: A

semantic approach for merging ﬂexibility with strict-

ness. pages 434–439.

Buyuksalih, I., Bayburt, S., Buyuksalih, G., Baskaraca,

A. P., Karim, H., and Rahman, A. A. (2017). 3d mod-

elling and visualization based on the unity game en-

gine – advantages and challenges. ISPRS Annals of

the Photogrammetry, Remote Sensing and Spatial In-

formation Sciences, IV-4/W4:161–166.

Cambridgeshire, C. (2025). Digital twins: Better use of

data. Accessed: 2025-02-21.

Castelli, G., Tognola, G., Campana, E. F., Cesta, A., Diez,

M., Padula, M., Ravazzani, P., Rinaldi, G., Savazzi, S.,

Spagnuolo, M., and Strambini, L. M. (2019). Urban

intelligence: a modular, fully integrated, and evolv-

ing model for cities digital twinning. 2019 IEEE 16th

International Conference on Smart Cities: Improving

Quality of Life Using ICT & IoT and AI (HONET-

ICT), pages 033–037.

Interconnecting Urban Networks: A Novel Approach to Digital Twins Through GlassBox Adaptation

203

CKAN (2025). Ckan - open source data portal. Accessed:

2025-02-21.

Ferr

e-Bigorra, J., Casals, M., and Gangolells, M. (2022).

The adoption of urban digital twins. Cities.

FIWARE Foundation (2024). Fiware: Open source platform

for smart applications. Accessed: 2024-11-27.

for BOLD Cities, C. (2025). Urban digital twins. Accessed:

2025-02-21.

Hardi, R., Nurmandi, A., Purwaningsih, T., and Ab-

dul Manaf, H. B. (2023). Exploration organiza-

tional interoperability in smart governance in indone-

sia and malaysia. In Stephanidis, C., Antona, M.,

Ntoa, S., and Salvendy, G., editors, HCI International

2023 Posters, pages 203–210, Cham. Springer Nature

Switzerland.

Hatzivasilis, G., Askoxylakis, I. G., Alexandris, G., Ani-

cic, D., Br

oring, A., Kulkarni, V., Fysarakis, K.,

and Spanoudakis, G. (2018). The interoperability of

things: Interoperable solutions as an enabler for iot

and web 3.0. 2018 IEEE 23rd International Workshop

on Computer Aided Modeling and Design of Commu-

nication Links and Networks (CAMAD), pages 1–7.

Karabulut, E., Pileggi, S. F., Groth, P., and Degeler, V.

(2024). Ontologies in digital twins: A systematic lit-

erature review. Future Generation Computer Systems,

153:442–456.

Kumar, S., Jeong, S., Ahn, I. Y., and Jarwar, M. A.

(2022). Things data interoperability through annotat-

ing onem2m resources for ngsi-ld entities. In 2022

IEEE International Conferences on Internet of Things

(iThings) and IEEE Green Computing & Communica-

tions (GreenCom) and IEEE Cyber, Physical & Social

Computing (CPSCom) and IEEE Smart Data (Smart-

Data) and IEEE Congress on Cybermatics (Cyber-

matics), pages 119–124.

Lu, Q., Parlikad, A. K., Woodall, P., Ranasinghe, G. D.,

Xie, X., Liang, Z., Konstantinou, E., Heaton, J., and

Schooling, J. (2020). Developing a digital twin at

building and city levels: Case study of west cam-

bridge campus. Journal of Management in Engineer-

ing, 36(3):05020004.

Maxis (2012). Inside the Glassbox (GDC 2012). Accessed:

2025-01-24.

Mihai, S., Yaqoob, M., Hung, D. V., Davis, W., Towakel,

P., Raza, M., Karamanoglu, M., Barn, B., Shetve, D.,

Prasad, R. V., Venkataraman, H., Trestian, R., and

Nguyen, H. X. (2022). Digital twins: A survey on

enabling technologies, challenges, trends and future

prospects. IEEE Communications Surveys & Tutori-

als, 24(4):2255–2291.

Mukherjee, T., Chander, D., Mondal, A., Dasgupta, K.,

Kumar, A., and Venkat, A. (2014). Cityzen: A

cost-effective city management system with incentive-

driven resident engagement. In 2014 IEEE 15th In-

ternational Conference on Mobile Data Management,

volume 1, pages 289–296.

Peldon, D., Banihashemi, S., LeNguyen, K., and Derrible,

S. (2024). Navigating urban complexity: The transfor-

mative role of digital twins in smart city development.

Sustainable Cities and Society, 111:105583.

Pliatsios, A., Kotis, K., and Goumopoulos, C. (2023).

A systematic review on semantic interoperability in

the ioe-enabled smart cities. Internet of Things,

22:100754.

Quek, H. Y., Sielker, F., Kraft, M., Akroyd, J., Bhave, A.,

Richthofen, A. V., Herthogs, P., Yamu, C., Wan, L.,

Nochta, T., Burgess, G., Lim, M. Q., Mosbach, S.,

and Balijepalli, V. S. K. M. (2021). The conundrum in

smart city governance: Interoperability and compati-

bility in an ever-growing digital ecosystem.

Rahman, H. and Hussain, M. I. (2020). A comprehen-

sive survey on semantic interoperability for internet

of things: State-of-the-art and research challenges.

Transactions on Emerging Telecommunications Tech-

nologies, 31(12):e3902.

Shih, C.-S. D., Yang, C.-M., and Cheng, Y.-C. (2015). Data

alignment for multiple temporal data streams without

synchronized clocks on iot fusion gateway. 2015 IEEE

International Conference on Data Science and Data

Intensive Systems, pages 667–674.

Tardieu, H. (2022). Role of gaia-x in the european data

space ecosystem. In Designing Data Spaces.

Thelen, A., Zhang, X., Fink, O., Lu, Y., Ghosh, S., Youn,

B. D., Todd, M. D., Mahadevan, S., Hu, C., and

Hu, Z. (2022). A comprehensive review of digital

twin — part 1: modeling and twinning

enabling technologies. Struct. Multidiscip. Optim.,

65(12).

Veltman, K. H. (2001). Syntactic and semantic interop-

erability: new approaches to knowledge and the se-

mantic web. New Review of Information Networking,

7(1):159–183.

Viola, F., Antoniazzi, F., Aguzzi, C., Kamienski, C., and

Rofﬁa, L. (2019). Mapping the ngsi-ld context model

on top of a sparql event processing architecture: Im-

plementation guidelines. In 2019 24th Conference of

Open Innovations Association (FRUCT), pages 493–

501.

Waddell, P., Garcia-Dorado, I., Maurer, S. M., Boeing, G.,

Gardner, M., Porter, E., and Aliaga, D. G. (2018). Ar-

chitecture for modular microsimulation of real estate

markets and transportation. ArXiv, abs/1807.01148.

SMEN 2025 - Special Session on Smart City and Smart Energy Networks

204