Next-Generation Smart Cities: An Overview and a Proposal for the Hub Architecture

Abstract

The smart city represents a new stage in urban evolution, driven by technological progress, social transformations, and the increasing emphasis placed on sustainability. This metamorphosis generates hub-type architectural models, used not only for data collection and interconnection but also for the management and monitoring of people, resources, and urban services. This discussion addresses how digital urbanism has followed different paths globally by synthesising technological, economic, social, and governance perspectives. Compared with traditional models of urbanisation, new smart cities are built not only for digital interconnection but also to be citizen-centred, environmentally friendly, and resilient to global crises. This article analyses recent scientific literature on the theoretical and practical foundations of technologies that support data-driven decision-making, infrastructure efficiency, and the delivery of inclusive public services. At the same time, major challenges are highlighted, such as the lack of system interoperability, information fragmentation, and the risks associated with excessive surveillance, which can generate social exclusion, as well as financial and political constraints. International examples from Helsinki, Barcelona, Dubai, and Singapore offer both models that have achieved success and critical lessons about the limits of these approaches. This paper is not limited only to the problems faced by smart cities. It also highlights the opportunities they can bring. Finally, based on the conclusions of the analysis carried out and the identified trends, a strategic framework is proposed, oriented towards responsible innovation, collaboration, and sustainability. This approach contributes to informing researchers, decision-makers, urban planners, and the public interested in the transformation of the urban environment.

1. Introduction

Throughout the past decades, urban areas have evolved to become complex socio-technical systems where millions of residents interact daily with infrastructures, digital services, and governance systems. Nonetheless, this has brought innovation possibilities but also significant sustainability, mobility, energy consumption, and quality-of-life challenges. Against these challenges, the concept of the smart city emerged, typically defined as combining high technologies with the governance of the city in an attempt to enhance efficiency, inclusiveness, and resilience [1].

Furthermore, the first wave of smart city initiatives was primarily focused on the deployment of Information and Communication Technologies (ICT) to digitalize the public sector and improve functional effectiveness. Traditional examples included online citizen portals, sensor-based traffic monitoring, and early experiments in open data platforms [2]. Though as much as these projects reflected concrete returns, they were developed piecemeal, leading to fragmentation, poor interoperability, and poor scalability. This isolated method also failed to foster citizen involvement and did not harness the power of data-driven decision-making at the city level.

As urban problems accumulate, the next-generation smart city vision moves away from isolated digital projects toward integrated ecosystems that connect citizens, processes, and infrastructures. Moreover, in the case of these ecosystems, interoperability as a brick-and-mortar factor becomes a requirement, making heterogeneous components ranging from Internet of Things (IoT) sensors and transportation systems to e-government portals seamlessly communicate with each other [3]. Moreover, information flows in real time, enabling predictive analytics, anticipatory service delivery, and responsive adaptation to live urban conditions. Citizens are no longer passive recipients of services but active participants, enabled to engage in decision-making processes through open and participatory platforms.

Therefore, emerging technologies frame this new era. The IoT extends the potential to collect high-grained data in the city environment, while artificial intelligence (AI) and machine learning (ML) provide the means to analyse it to extract useful insights. Consequently, blockchain technology presents options for decentralised governance, secure transfers, and tamper-proof audit trails, enhancing the chances of enhancing transparency and trustworthiness in public administration. Concurrently, technical advancements in 5G and edge computing lower latency and support distributed intelligence close to where data is created [4]. Quantum-inspired optimisation and quantum computing will make possible the combinatorial complexity of mobility planning, energy distribution, and large-scale simulation in the future.

From a new perspective on how we view things, we can say that Transit-Oriented Development (TOD) emerges as a relevant conceptual and operational framework for next-generation smart cities. Traditionally associated with land-use and transport planning, TOD is increasingly being reconceptualised as a data-centred layer that integrates public transport infrastructure, patterns of human mobility, traffic dynamics, pedestrian flows, and safety-related information. By structuring urban environments around high-capacity transport nodes, TOD provides a natural foundation for mobility-oriented, human-centred urban systems [5].

When combined with large-scale mobility data and machine learning techniques, TOD-related information supports real-time perception and predictive analytics, enabling demand forecasting, congestion mitigation, and risk-aware traffic and infrastructure management. Recent high-impact studies demonstrate the effectiveness of data-driven and machine learning-based approaches in transit-oriented contexts, highlighting their contribution to intelligent traffic safety analysis and adaptive urban mobility systems [6].

Within this paradigm, TOD can be understood as a foundational data layer for smart mobility ecosystems, facilitating human–infrastructure interaction and supporting adaptive decision-making. The hub architecture proposed in this paper is positioned within this TOD-supported framework, leveraging transit-centric data flows to enhance scalability, responsiveness, and urban safety in next-generation smart cities.

All this aside, there are still significant obstacles to be overcome. Governance systems must evolve to enable the integration of technology solutions while not compromising inclusiveness and equity [7]. Furthermore, security and privacy concerns are compounded with the potential of networked devices as well as sensitive citizen information. Nevertheless, additionally, sustainable financing and regulatory alignment across municipality, national government, and global levels need to be guaranteed to facilitate long-term take-up and effect. Nonetheless, these demands highlight that smart cities cannot be viewed as technological rollouts but as socio-technical systems within which institutional ability, civic participation, and cultural settings all have considerable roles to play. In this context, technical and practical discourse regarding smart cities increasingly implies the need for coordinated reference architectures and hub designs. These are abstract models that can guide implementation in various contexts and enable interoperability, scalability, and extensibility.

Therefore, through the coordination of technology design with policy objectives and user needs, such architectures can enable cities to evolve from disjointed initiatives to integrated ecosystems.

The aim of this paper is to explore these trends, provide a structured overview of future paradigms for smart cities, and examine how hub-based structures can overcome present shortcomings. The argument will be based on the existing literature, critically assessing trends and deficiencies, before addressing methodological issues and presenting an exemplary proposal for a smart city-oriented infrastructural framework as a development concept.

The objective of this research is to carry out a detailed analysis of the specialised literature on current trends at the level of contemporary cities and to delineate a set of models and strategies for designing the transition towards smart cities. Based on the research conducted in the field, this paper proposes, at a conceptual level, the development of a prototype platform aimed at supporting the digitalisation of current processes within smart city governance. In this way, relevant strategies will be outlined and proposed, which can strengthen the gradual transition from today’s cities to those of the future, where citizens will have easy access to all the advantages and opportunities of a developed city.

This paper is divided into six sections, which present the subject approached in stages:

Section 1 is represented by the introduction and illustrates the current level of smart cities.

Section 2 focuses on the detailed analysis of the specialised literature in the field.

Section 3 presents the materials and methods approached to expose the theoretical way of implementing an intelligent system dedicated to the community and integrated in cities.

Section 4 illustrates a conceptual proposal of a smart city.

Section 5 analyses the presentation of a conceptual prototype for smart cities, showing how this type of system could be used if it were to be developed.

Section 6 presents the opportunities and future directions of development.

And the paper ends with conclusions.

2. Review of the Specialised Literature

2.1. Review Methodology

This section follows a structured narrative review approach to synthesise recent research and guidelines concerning next-generation smart cities and reference architectures based on integrated systems or subsystems.

Sources were identified through targeted searches in relevant scientific databases and indexing services, as well as by tracing citations within the specialist literature. The search strings combined terms related to smart cities, interoperability, reference architectures, data platforms, IoT/edge/5G, AI/ML, and governance.

To preserve thematic relevance, priority was given to research that proposes an actual implementation in the field of future smart cities. Based on the detailed analyses carried out, this research outlines the directions for designing a prototype generically referred to as a Smart City Hub. This paper does not aim to develop the prototype at the implementation level but is instead confined to an analysis of the most recent studies in the field and to the identification of the most highly regarded methods, techniques, and strategies for the conceptual design of such a system.

The uniqueness of this paper lies in the breadth of application areas covered by the conceptual system proposed at the prototype level and, at the same time, in the fact that it addresses the most efficient methods in relation to the research conducted in the field. Since the body of information identified in existing studies covers only isolated sectors of applicability without extending the overall scope of the operational approach, the adopted strategy is to propose a system which, through its capabilities, can fulfil multiple attributes and can be classified as a hub by virtue of its ability to collect and interconnect several operational sectors within everyday urban environments.

Each selected research source was examined using a common set of analytical dimensions:

-

Urban domain or domains addressed;

-

Scope of integration;

-

Interoperability approach;

-

Type of evaluation (concept, simulation, testbed, or implementation);

-

Metrics or isolated proposals only;

-

Considerations regarding data governance, privacy, and security.

This approach enables systematic comparison and informs the reader about existing research gaps and the derived design requirements presented at the end of this paper.

2.2. Thematic Synthesis of the Literature

To make the review explicit and comparable, the literature is discussed at an analytical level and identifies the weaknesses encountered within the proposals presented in the studied works. In this way, a synthesis is achieved and a set of gaps in the reviewed research is highlighted, and, subsequently, based on these derivations, the foundations of the proposed conceptual hub prototype are identified.

To ensure a rigorous and transparent analysis of the existing studies, the literature selection process followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 methodology. Systematic searches were conducted in the main scientific databases (Scopus, IEEE Xplore, MDPI, and ScienceDirect) and were subsequently complemented by web sources and additional bibliographic references.

The stages of identification, screening, eligibility, and inclusion were documented and are illustrated in Figure 1.

Over roughly two decades, “smart city” research has shifted from a technology-led view (networks, sensors, platforms) to a more systemic understanding that blends technology with governance, participation, and sustainability. Moreover, early deployments demonstrated tangible benefits but also exposed fragmentation and interoperability gaps, prompting frameworks that stress integrated, citizen-involving ecosystems supported by common architectures and standards.

2.2.1. Dimensions and Conceptual Frameworks

In research [9], the author Townsend presents a different perspective on the subject and places smart cities at the intersection of rapid urbanisation and ubiquitous computing, advocating “civic” approaches that prioritise participation, openness, model transparency, and local adaptation over monolithic, vendor-driven systems. In addition, he highlights open data, civic hacking, and resilience-by-design principles as essential for avoiding technological lock-in and exclusion.

In [10], a data-driven approach is proposed for identifying “potential TOD areas”, defined as spaces that already exhibit physical characteristics specific to TOD but have limited public transport connectivity and can be prioritised for planning interventions. The study develops a Random Forest-mediated TOD Potential Identification (RF-TPI) model applied in Hong Kong, in which the city is discretised into grids and built environment indicators are derived from multi-source urban data. The results indicate good performance, highlighting the usefulness of the method for selecting TOD locations.

This literature review clarifies the contested term “smart city” and identifies core dimensions used to characterise urban smartness (e.g., economy, people, governance, mobility, environment, living), alongside metrics and performance/initiative mappings. It underscores definitional ambiguity, proposes a structured dimensional view, and surveys policy/institutional documents to anchor practice [11].

Providing one of the most cited operational definitions, linking smartness to investments in human/social capital and ICT to drive sustainable economic growth and quality of life, with participatory governance as an enabler, the work positions ICT as a strategic factor within a broader bundle of urban production factors [12].

Hence, it introduces an eight-factor framework, management and organisation, technology, governance, policy context, people and communities, economy, built infrastructure, and natural environment, to consequently analyse and design smart-city initiatives. Nevertheless, the framework addresses the earlier piecemeal orientation by emphasising interdependencies across institutional, social, and technical subsystems [13].

2.2.2. Governance, Inclusion, and Ethical Policy

A relevant example is presented in [14], which highlights the need for a sustainable, inclusive and people-centred urban digital transformation. It provides cities with assessment tools, roadmaps, technical support, case studies, model policies and concrete measures for implementing policies that do not exclude any citizen. At the same time, the framework links digital governance with citizens’ rights, institutional capacity and public participation, going beyond a vision limited to simple technological implementation.

In [15], the relationship between the economic performance of TOD and social equity in areas surrounding railway stations is analysed, using multi-source data and predictive modelling. The authors extend the node–place framework to the node–place–economic (NPE) model, integrating the dimensions of transport, urban development and economic externalities. The study, applied in Dalian (China) across 46 stations, defines the analysis area through a 20-min walking buffer. The results indicate high model performance (Mean Squared Error (MSE) ≈ 0.00092, R² ≈ 0.916 for training; MSE ≈ 0.00113, R² ≈ 0.884 for testing), suggesting a good capacity to predict the social equity index.

Furthermore, in [16] it is shown how more than 100 European cities pursue intelligent, sustainable and socially responsible development, using advanced technologies. The work highlights objectives such as the green transition, replicable solutions, the involvement of local ecosystems and the connection with the European Union’s mission on climate neutrality by 2030. At the same time, the document emphasises platforms for smart solutions, open data sharing, common investment mechanisms and urban laboratories dedicated to testing and experimentation. Thus, cities are not only consumers of technology but become hubs of social and economic innovation and experimentation.

Another relevant example presented in [17] highlights the importance of a human-centred approach and ethical practices in the development of smart cities. The research shows that, although current initiatives pursue urban digital transformation for efficiency and sustainability, regulatory frameworks remain fragmented, and adaptive governance and levels of inclusion vary significantly across cities. The authors propose a set of best practices that integrate ethical principles, equitable access, and sustainable policies into smart city projects, emphasising community engagement, data transparency, and the adaptability of urban policies.

A further review of the literature examines the major challenges associated with inclusive governance in smart cities, highlighting how digital inequalities and barriers to access to technology can lead to social exclusion. Study [18] provides a reflective framework for understanding how included or excluded actors influence urban governance and emphasises the need for policies that ensure public value and equal benefits for all citizens, thereby contributing to the orientation of urban digital policy design towards genuine inclusion.

2.2.3. Standardisation and Performance Measurement

With regard to standardisation, ref. [19] offers globally recognised indicators for urban services, quality of life, smart cities and resilience. Moreover, it establishes core and supporting indicators, widely used by municipalities through the World Council on City Data for comparable reporting. According to the guideline, these standards are designed to work together, providing a comprehensive overview of urban performance, and their application is a prerequisite for implementation.

A study from the specialist literature proposes an indicator framework for evaluating smart city initiatives, emphasising the importance of performance measurement in supporting strategic decision-making. The analysis highlights the use of indicator systems at the international level to assess progress across the planning, implementation, and monitoring stages, as well as the need for a coherent set of key performance indicators (KPIs) that enable comparability and the monitoring of smart urban development [20].

The significant work [21] in the field addresses the development of a European system for measuring the performance of smart cities, presenting the process of designing a performance indicator framework used to set strategic objectives and monitor progress. The described project highlights the benefits of a transparent and structured KPI system that facilitates cooperation between cities, comparability of results, and the continuous improvement of smart urban strategies.

2.2.4. Enabling Technologies and City Data Platforms (IoT, AI/ML, Edge, Blockchain)

The work [22] presents an analytical framework for sensor- and Big Data-based IoT applications aimed at environmental sustainability. The article synthesises the state of IoT applications (environmental monitoring, energy efficiency, operational optimisation) and discusses challenges such as data quality, platform integration, privacy and security, as well as issues of data governance.

In [23], a predictive approach based on machine learning is proposed for estimating metro ridership as support for TOD planning, correlating transport demand with land use around stations. It uses time series collected from 38 stations and characterises the built environment within an 800 m catchment area. Twelve “white-box” and “black-box” models are compared, with hyperparameter tuning performed through grid search and Bayesian optimisation, using 10-fold cross-validation. The results indicate good performance for certain “base learners”, while ensemble-type models improve prediction accuracy.

Paper [24] proposes an approach based on a Heterogeneous Graph Neural Network (HGNN) for classifying urban functional regions, showing that the integration of population mobility with data on POIs and building morphology significantly improves the understanding of urban functions and supports public transport-oriented planning decisions. A heterogeneous graph is constructed in which urban blocks and metro stations are nodes, and the links between them are defined both through spatial proximity and through functional interactions reflected by passenger flows, allowing the model to simultaneously capture “geographical” relationships and “mobility” relationships. In experiments conducted in Beijing, the method achieves OA = 81.56%, F1Macro = 77.09% and Kappa = 70.03% using only 20% labelled data for training/validation, indicating efficiency under conditions of limited labelled data. The authors show that the inclusion of metro flow data increases classification accuracy by 6.87%, demonstrating the decisive role of mobility in identifying urban functions and in better distinguishing commercial/office areas from residential ones.

In [25], a comprehensive review of the concept of smart urban governance is presented, arguing that “smart city governance” involves building new forms of human collaboration supported by ICT, with the aim of achieving more efficient outcomes and more open and transparent decision-making processes. The authors emphasise the fragmented nature of existing approaches, highlighting that many initiatives are developed in isolation and without real integration at the institutional level. They also underline the political and social nature of socio-technical visions and propose analysing smart urban governance as a complex institutional process that involves multiple levels of decision-making and social actors, rather than simply as a technological issue.

The research in [26] situates smart cities within the logic of inclusive growth and shows how smart policies must reduce inequalities, protect vulnerable groups and create more open mechanisms of local governance. The document and the accompanying report emphasise that data platforms, open data practices and evidence-based recommendations can support cities in post-crisis recovery and in connecting with climate objectives.

The work [27] proposes policy benchmarks for the ethical governance of smart city programmes, covering five fundamental areas: ICT accessibility, privacy impact assessment, responsibility for cyber security, digital infrastructure and open data. It is based on questionnaires and interviews and identifies policy gaps, offering a comparative framework for aligning standards and practices.

Nonetheless, article [28] reviews the urban potential of AI and Big Data and proposes a framework that links technology to cities, insisting on the integration of dimensions such as culture, urban metabolism and governance for alignment with the Sustainable Development Goal and the New Urban Agenda. Furthermore, the work is a synthesis that maps the use of AI/IoT in cities and argues for approaches that avoid monolithic solutions and favour integration into broader urban systems.

2.2.5. Cross-Domain Applications and Intelligent Systems Approaches

The paper in [29] addresses the use of blockchain in personal document archiving services. Public information indicates that the authors managed to achieve an integrated system for the archiving of personal documents, also offering increased data security. Based on the bibliographic records, the archiving solution harnessed through the properties of blockchain for the integrity and traceability of documents has great success due to the benefits it brings.

The research [30] brings to the forefront a system that functions as a hub of resources (webinars, reports, technical materials), initiatives and events dedicated to smart cities, with sections for the community, ambassadors and conferences. It is a site that highlights the necessary steps to support municipalities in the planning and implementation of smart city projects and aggregates links to event calendars and the resource centre.

A study proposes a hybrid model based on machine learning for the dynamic classification of railway/metro stations, highlighting the role of correct typologisation in data-driven TOD strategies. The study extends the classical node–place paradigm through the Node–Place–Ridership–Time (NPRT) framework, which integrates station connectivity, the built environment, ridership and temporal variation in order to capture the dynamics of use over time. Methodologically, regression models (Multiple Linear Regression (MLR), Deep Neural Network (DNN), K-Nearest Neighbours (KNN)) are combined for ridership forecasting with unsupervised clustering techniques (K-Means Clustering (K-Means), Agglomerative Nesting (AGNES), Density-Based Spatial Clustering of Applications with Noise (DBSCAN), Gaussian Mixture Model (GMM)) for identifying station typologies. The results indicate high performance (MSE < 0.012), while the neural network used for the final classification achieves 100% accuracy across seven time intervals and 98.15% in the eighth. The study demonstrates the potential of a data-driven approach for TOD planning, facilitating the prioritisation of investments and their adaptation to the dynamics of each station [31].

In the research in [32], an intelligent system is presented that uses blockchain in the context of medical diagnosis. From the public metadata there results a focus on the integration of blockchain for data security/integrity and on the system architecture. We can consider that this research brings a beneficial point in the process of developing a smart city due to the methods approached for the digitalisation of the current medical system.

Furthermore, the work [33] addresses the design of a medical diagnostic system based on the integration of blockchain and AI technologies. Nonetheless, the authors investigate how AI algorithms can analyse medical data to support doctors in establishing a diagnosis, while blockchain is used to guarantee the security, traceability and integrity of medical information. Nevertheless, the main contribution consists in defining an architecture that combines the automatic analysis of clinical data with the decentralised and immutable storage of diagnostic results. Nonetheless, the paper has an abstract character, without presenting details regarding the types of patients’ medical data used for analysis. It constitutes a good way of developing the medical sector within the smart city system.

Moreover, in a complementary direction, in [34], an architectural model is proposed for the verification and identification of inappropriate products through the use of IoT, blockchain and artificial intelligence. Nonetheless, the system aims to provide a mechanism through which consumers can validate the authenticity and quality of products. Consequently, IoT sensors are used for collecting product data, AI algorithms process and classify the information, and blockchain ensures the traceability and security of the recorded data. Nonetheless, the study describes the structure of the architecture and the way in which the components interact, and it is a suitable system for the development of intelligent systems. Another field addressed for intelligent systems, which can be integrated into a system for future cities, is image-based medical diagnosis. The authors examine the possibility of early diagnosis of ocular conditions through the application of artificial intelligence to ophthalmological images. The paper explores the use of image pre-processing techniques and the training of deep learning models to detect early pathological changes, such as those associated with glaucoma or macular degeneration. Hence, the advantage of this approach lies in the early identification of conditions, which could increase the chances of effective treatment. However, the authors highlight the challenges related to the quality and diversity of available imaging data, the risk of bias in the training sets and the need for rigorous clinical validation before large-scale implementation [35].

In [36], the focus is on a different field but with a methodology relevant for intelligent systems, monitoring and controlling energy consumption, which, within a smart city, is a good subject to address. The paper proposes an intelligent system capable of analysing data on energy consumption and deciding in real time on its regulation, with the aim of reducing costs and optimising efficiency. The authors use fuzzy logic techniques to deal with uncertainty and variability in data, reflecting the way in which human decisions can be imitated in automated processes. Although the solution is applied in the energy context, the methodology based on IoT sensors and decision algorithms can also be transferred to other fields, including the medical one.

Consequently, in a less technological and more geopolitical register, the work [37] proposes an intelligent system for the analysis of the level of gain of development hegemony for great powers. It studies statistical and economic indicators in order to comparatively evaluate the degree of development and the geopolitical influence of states. The use of intelligent algorithms for data correlation and interpretation is suggested, but the methodology remains at a conceptual level, and the practical applicability is limited to political analysis. Hence, the relevance for the field of intelligent technologies lies rather in the demonstration that data processing methods and AI-assisted analysis can also be used in socio-political contexts.

In

Table 1. Synthesis of the most relevant research in smart cities.

No.	Title	Technologies Used	Description	Results	Limitations
[38]	Replication of Smart-City Internet of Things Assets in a Municipal Deployment	A standardised hardware (HW)/software (SW) and radio frequency (RF) set for the IoT, together with cyber-security practices.	It is a standardised framework for the design, operation and replication of IoT assets for smart cities. It is implemented and tested in a municipal testbed and then re-replicated.	A large number of sensor data records collected over eight months. Demonstrates inter-jurisdictional replicability. Technical and organisational barriers are inventoried, along with ways to address them.	The paper is focused on its non-modular architecture. It does not track operational key performance indicators in a specific urban domain.
[39]	LoRaWAN for Smart City IoT Deployments: A Long Term Evaluation	Long Range Wide Area Network (LoRaWAN), commercial and custom gateways (Kerlink iBST, Raspberry Pi (RPi) + iC880A), air-quality sensors, The Things Network (TTN) and Message Queuing Telemetry Transport (MQTT).	A city-wide LoRaWAN network in Southampton for connecting data from air-quality sensors. It measures delivery rates and end-to-end delays. Mixed topology of gateways placed randomly.	More than 135,000 messages from 20 devices. A 99% rate of received messages arriving in under 10 s. City-scale coverage is confirmed, and the cost/benefit between do-it-yourself (DIY) and commercial gateways is discussed.	Sub-optimal placements reduce “data completeness”. Long-term sustainability is not guaranteed with LoRaWAN.
[40]	From Data to Action: Exploring AI and IoT-driven Solutions for Smarter Cities	IoT, AI and ML.	Data-driven urban management system with three use cases: road safety, energy efficiency, and early detection of potholes, floods, and fires. It also includes a case study carried out in Aveiro.	The system demonstrates the generation of actionable insights for safety, efficiency, and sustainability in real cities.	Preprint focused solely on system description and case study. The paper does not provide quantitative city-level performance indicators.
[41]	Smart City IoT Services Creation through LargeScale Collaboration	Future Internet Ware (FIWARE), Next Generation Service Interface (NGSI)/NGSI Linked Data (NGSI-LD), Service-Oriented Architecture (SOA).	Methodology for the rapid creation of smart-city services through componentisation and pan-European collaboration (SynchroniCity project). Defines 15 robust services and demonstrates their reuse across cities.	Includes 35 pilot services in 27 cities and, in total, 38 active operational service instances. Provides concrete implementations in Santander and in three cities in Denmark.	Focus on interoperability and reuse. This work does not address city-specific urban impact results.
[42]	Modeling Trees For Virtual Singapore: From Data Acquisition To CityGML Models	Light Detection and Ranging (LiDAR) and satellite imagery. Procedural three-dimensional (3D) modelling and City Geography Markup Language (CityGML).	Framework for populating Virtual Singapore with urban trees represented semantically and biologically: data acquisition (LiDAR/satellite), tree extraction and quantification, multi-LOD 3D modelling in CityGML.	Preliminary results and pipeline challenges relevant for vegetation modelling in the city’s digital twin.	Subsystem-level paper, not covering complete urban services. Highlights difficulties in data acquisition and modelling.
[43]	Estimating Smart City sensors data generation	Analysis of the sensor layer in Barcelona, smart-city architectural layers based on sensing, network, middleware, and applications (apps).	Evaluation of the city of Barcelona with a focus on data collection. Estimates the daily volumes generated by sensors and projects a full-coverage scenario. Proposes solutions for reducing data traffic and improving management.	Provides quantitative estimates for data generation and recommendations for efficient city-scale management.	Estimation study does not include live measurements across all data streams. System access is restricted due to hardware limitations.
[44]	International Case Studies of Smart Cities	Integrated ubiquitous city (u-city) platform with u-portal, u-facility management, u-security, u-disaster, u-environment. Integrated operations centre and media kiosks.	Case study by the Korea Research Institute for Human Settlements (KRIHS) and the Inter-American Development Bank (IDB): Pangyo built as a smart city from the greenfield stage. One-shot implementation in 3–4 years. Secured 75 million United States dollars (USD) in allocated development gains. Provides integrated services through a central command centre.	Large-scale implementation with multiple categories of public services. Rapid approach of constructing the city and smart-city systems in parallel.	Report oriented towards description and policy, with few detailed operational metrics.
[45]	Aveiro Tech City Living Lab: A Communication, Sensing and Computing Platform for City Environments	Fibre network and millimetre wave (mmWave) links, Wireless Fidelity (Wi-Fi), Intelligent Transport Systems G5 (ITS-G5), Cellular Vehicle-to-Everything (C-V2X), Fifth-Generation mobile network (5G), Long Range Wide Area Network (LoRaWAN), sensors (Radio Detection and Ranging radar, Light Detection and Ranging LiDAR, video, environmental), edge computing (Edge), cloud computing (Cloud) and a data platform.	Describes the architecture, implementation, and operation of the Aveiro Tech City Living Lab (ATCLL) in Aveiro. Connects fixed points and public transport. Collects and processes data for the Internet of Things (IoT) and Intelligent Transport Systems (ITS).	Large-scale multi-protocol infrastructure with real mobility and safety use cases. Provides field-measured indicators.	Article focused on infrastructure and capabilities. Does not report long-term public policy KPIs.
[46]	SmartSantander: IoT experimentation over a smart city testbed	Large-scale IoT testbed. Multiple federated urban scenarios. Architecture for experimentation. Support for smart-city applications.	Presents the architecture and implementation stage of the testbed. Concentrates a large number of IoT devices in real scenarios, federated into a single testbed for large-scale experimentation.	Reference platform used for experimentation and evaluation of IoT concepts under real-world conditions. Contributions to Future Internet architectures.	Focus on research and experimentation infrastructure, not on operational services or urban KPIs.
[47]	Romanian Municipalities and the Smart city Framework	Electronic government (e-government), mobile (Tranzy), smart street, public Wireless Fidelity (Wi-Fi), waste sensors, open data.	Comparative analysis of the most notable projects implemented in Cluj-Napoca, Iași, and Alba Iulia, across six components: governance, quality of life, citizen, environment, infrastructure/mobility, and economy. Includes site visits and municipal sources.	Ranking of projects in 2023: Cluj-Napoca 63, Iași 56, Alba Iulia 49. Covers implementations in the areas of transport, e-services, and environment.	Smart initiatives are still in early stages. Public information is limited. Some projects are delayed or unfinished.
[48]	IoT in urban development: insights into smart city applications, case studies, challenges, and future prospects	IoT, sensors, AI, cloud computing, data analytics, wireless communication technologies (WiFi, Bluetooth, RFID, cellular 4G/5G, NB-IoT), edge computing.	Comprehensive analysis of IoT-based smart city ecosystems. It analyses IoT components, integration with AI and networking technologies, and reviews several real-world smart city case studies. It investigates adoption challenges, particularly aspects related to security, privacy, interoperability, and governance.	IoT enables efficiency improvements in transportation, energy management, waste management, healthcare, and infrastructure monitoring. It provides examples of successful IoT implementations in cities and highlights best practices and cross-domain synergies.	The study is predominantly based on reviews and case studies. It does not provide KPIs, experimental benchmarks, or real-time validation through pilot deployments.
[49]	Digital twin technology in smart cities: a step towards intelligent urban management	IoT, AI, Big Data analytics, cloud computing, edge computing, 5G connectivity, blockchain-based security, predictive analytics.	A review that analyses the concepts, architectures, and technologies underpinning urban digital twins. It proposes a functional pipeline and a Digital Twin Implementation Readiness Level (DT-IRL) to align technical requirements with the real needs of cities.	It provides a structured multi-domain synthesis and defines clear directions for large-scale urban implementations. It highlights the role of IoT–AI–edge–cloud integration in achieving real-time performance.	Without proprietary implementation or direct experimental validation. It analyses existing pilot projects or case studies. Comparative quantitative evaluations between real platforms and operational outcomes are lacking.
[50]	SWOT analysis of smart city frameworks presented as a systematic literature review	ICT, IoT, data platforms, governance frameworks, environmental monitoring technologies.	Systematic literature review combined with SWOT analysis of existing smart city frameworks, focusing on core components rather than implementation details. The frameworks are classified into human-centred, technology-centred, human–technology integrated, and environment-focused categories.	It identifies four essential dimensions: technological infrastructure, governance, citizen engagement, and environmental sustainability. It indicates cybersecurity risks and technological dependency, while good governance and citizen inclusion represent key factors for sustainable smart cities.	Conceptual and analytical study. Without experimental validation, quantitative KPIs, or real-world performance measurements. The results depend on the quality and scope of the analysed literature and may be subject to a degree of subjective interpretation.
[51]	Smart cities and urban resilience: evaluating the impact on emergency response in China	IoT, Big Data analytics, cloud computing, smart city platforms, urban data integration systems.	Empirical study investigating the impact of smart city construction (SMC) on the high-quality development of urban emergency response in China. It employs a quasi-experimental design using Difference-in-Differences (DID) and PSM-DID methods applied to panel data from 237 prefecture-level cities.	The results show that smart city development has a significant positive effect on the quality of emergency response. Urban resilience is confirmed as a key mediating factor. The analysis indicates stronger effects in eastern regions compared to western regions.	The study is limited to Chinese prefecture-level cities, which reduces the generalisability of the findings to other national contexts. It focuses on policy-level impacts rather than on specific smart city technologies.
[52]	A methodological framework for the selection of key performance indicators for assessing smart city solutions	Urban innovation platforms, smart energy systems, ICT platforms, data analytics, KPI frameworks (CITYkeys, SCIS), stakeholder-oriented evaluation models.	The paper proposes a holistic methodological framework for the selection and structuring of KPIs intended to assess smart city solutions. The framework is built in six steps and integrates technical, environmental, economic, social, ICT, and legal perspectives, being citizen-centred.	The implementation of the framework led to the development of a repository of 75 KPIs organised across six dimensions and correlated with evaluation levels (building, district, city). The framework was validated in European demonstrator projects and demonstrates high applicability for monitoring.	The framework is oriented towards evaluation and monitoring, without providing operational values or direct results regarding urban performance. Full implementation requires extensive data access and a high level of institutional involvement.
[53]	Smart city indicators and the conceptual problems of measuring smart cities	Smart city indicator frameworks, composite indicators, proxy indicators, urban statistical data, monitoring and decision-support tools.	Conceptual and methodological paper that critically analyses the use of indicators for measuring urban performance. It examines the gap between the concept of a smart city and the actual content of the indicators used in rankings and evaluations, demonstrating that many measures are not linked to technology or digitalisation.	It identifies major issues of validity, relevance, and comparability of the indicators used in smart city assessments. It demonstrates that composite indicators can lead to erroneous interpretations and misleading city rankings. It proposes a return to individual indicators clearly correlated with digitalisation.	Predominantly theoretical work, without validation or application through real-time testing. It does not propose a new set of KPIs and does not provide evaluations of urban performance, but rather a critical analysis of existing practices.
[54]	Smart cities and communities: a key performance indicators framework	IoT, cyber-physical systems (CPS), sensors and actuators, wired and wireless networks (Wi-Fi, LPWAN, 5G, fibre), data platforms, data analytics, cloud and edge computing, interoperability frameworks.	Technical report proposing the Holistic Key Performance Indicators (H-KPI) Framework for the evaluation of smart cities and communities. The methodology adopts a multi-level approach and integrates alignment with community priorities, investment efficiency, and data flows.	It introduces five metrics: alignment of KPIs with community priorities, investment alignment, investment efficiency, information density, and quality of benefits. It provides a structured analytical method, support for comparability between cities, and an assessment of the maturity of urban systems.	Methodological and standardisation framework, without practical implementation or empirical results. It requires large volumes of data, institutional capacity, and adaptation to the local context. It does not provide predefined values, but rather a selection mechanism.
[55]	Edge AI for smart cities: foundations, challenges, and opportunities	Edge AI, IoT, CPS, AI/ML, edge computing, cloud computing, heterogeneous sensors, 5G/next-generation networks, embedded edge devices (Raspberry Pi, NVIDIA Jetson, microcontrollers), learning model optimisation (compression, pruning, federated learning).	Survey article that provides an analysis of Edge AI architectures for smart cities. It proposes a unified taxonomic view of key components: urban applications, sensor data, learning models, and infrastructure. The domains of transport, healthcare, buildings, environment, and manufacturing are analysed.	It offers a synthesis of Edge AI applications in smart cities and highlights patterns, recurring challenges, and strategies. It contributes an integrated conceptual framework that can guide research, design, and public policies in urban contexts.	It has no implementation. It does not provide urban performance indicators or comparative evaluations between cities. Validation is indirect, based on the analysed literature, and applicability depends on future studies and real-world implementations.
[56]	Integrating smart city technologies and urban resilience: a systematic review and a research agenda for urban planning and design	IoT, Geographic Information Systems (GIS), AI, digital twins, Big Data analytics, blockchain, open data platforms, participatory digital platforms.	Systematic review (PRISMA 2020) of 115 studies analysing how smart city technologies are integrated into urban planning tools, governance, and social processes to support urban resilience.	It demonstrates that smart technologies contribute to urban resilience. It identifies IoT and GIS as the technologies with the strongest foundation for monitoring, early warning, and operational management, while AI, digital twins, and blockchain largely remain at pilot or conceptual stages.	Limited to academic literature and to cities predominantly from high-income countries. There is limited comparative evidence regarding the long-term impact of emerging technologies. The analysis highlights imbalances.
[57]	Explainable Artificial Intelligence for smart city applications: a secure and trusted platform	AI, ML, Deep Learning (DL), Explainable Artificial Intelligence (XAI), IoT, cyber-physical systems, cloud computing, edge computing, cybersecurity frameworks.	Analytical chapter investigating the role of XAI in smart cities. It discusses the transition from “black-box” AI models to “white-box” approaches, highlighting requirements for interpretability, transparency, and explainability to enhance security.	It demonstrates that XAI can improve trust, control, and security in intelligent systems by facilitating the understanding of automated decisions and the detection of biases and vulnerabilities.	Synthesis work, without implementation or experimental evaluation. Performance indicators or comparative studies between cities are not discussed. Applicability depends on future developments.
[58]	Artificial intelligence of things for a sustainable smart city brain and digital twin systems: pioneering environmental synergies between real-time management and predictive planning	Artificial Intelligence of Things (AIoT), IoT, AI, ML, Big Data analytics, Urban Brain (UB), Urban Digital Twin (UDT), Cyber-Physical Systems of Systems (CPSoS), cloud and edge computing.	Conceptual framework that integrates UB/UDT platforms into an AIoT-based cyber-physical architecture. It aims to link real-time urban management with long-term predictive planning, supporting sustainable operations and decision-making processes in smart cities.	Bidirectional integration between operational analytics and predictive simulation to support urban sustainability. It highlights the potential of AIoT to improve environmental performance, resilience, and resource efficiency in smart cities.	Conceptual and theoretical study, without practical implementation or experimental validation.
[59]	Integration of IoT and Digital Twin for intelligent management of urban underground pipe galleries in smart cities	IoT, DT technology, cloud computing, edge computing (Raspberry Pi), ML, predictive analytics, LoRaWAN communication, RESTful APIs.	Intelligent management system for urban underground pipe galleries through the integration of IoT-based real-time monitoring with a DT-based virtual model. The system enables continuous data acquisition, predictive maintenance, and optimisation of underground infrastructure operations.	The evaluation demonstrates a prediction accuracy of 92% in detecting potential failures. Maintenance costs are reduced by 35%, while system downtime is decreased by 40%.	Validation is carried out in a simulated and small-scale environment. Challenges remain for large-scale deployment, including scalability, data volume management, network latency, and cybersecurity risks.
[60]	What indicators to use, why and when? (Evaluation of Smart and Sustainable City Development)	Smart and sustainable city indicator frameworks, urban performance indicators, evaluation and monitoring tools, statistical and qualitative analysis methods.	Research that proposes a taxonomy and a structured process for indicator selection in smart and sustainable city evaluations. The study analyses over 1500 indicators, comparing smart city indicators and sustainable city indicators according to their focus.	It demonstrates differences between indicators for smart cities and for sustainable cities. Smart city indicators focus on short-term technological efficiency, while sustainability indicators emphasise long-term impacts.	Methodological research without experimental implementation. The results depend on the analysed indicators and literature. Applicability requires local adaptation and expert interpretation.
[61]	Smart cities and IoT: an in-depth analysis of global research trends and future directions	IoT, AI, Big Data analytics, cloud computing, edge/fog computing, Natural Language Processing (NLP), Latent Semantic Analysis (LSA), TF-IDF, K-means clustering, Scopus database.	It analyses trends in the field of smart cities and IoT. The study processes 14,309 Scopus-indexed publications (2010–2024) using text-mining and LSA techniques to identify dominant research areas, emerging topics, countries, authors, and influential journals.	It identifies ten major research clusters, including security and privacy, sustainable smart cities, smart healthcare, smart governance, smart transport, cyber-physical systems, Industry 4.0, IoT in energy, and smart supply chains.	Exclusive review without implementation or validation. The results depend on the coverage of the Scopus database and the text-mining methods used.
[62]	Smart city technologies for sustainable urban planning: evidence and equity lessons from Shenzhen	IoT, AI, Big Data analytics, smart energy grids, renewable energy systems, AI-based traffic management, sponge city infrastructure, vertical greening, cloud and edge computing.	Study analysing the integration of smart city technologies into urban planning in Shenzhen. Four initiatives are evaluated: smart and renewable energy grids, AI-based traffic management, sponge city infrastructure, and vertical greening solutions.	Reduction in energy consumption by 15% (≈1.6 TWh/year), reduction in CO2 emissions by 20% through AI-based traffic management, reduction of flood incidents by up to 60% in pilot areas, expansion of urban green spaces by 30%, and improvement in air quality.	High implementation and maintenance costs, social equity issues, risks related to data security and privacy, algorithmic bias in AI traffic systems, and limited generalisability.
[63]	Review and analysis of smart city evaluation frameworks	ICT platforms, smart city evaluation frameworks, indicator-based evaluation systems, urban data, benchmarking tools.	Analytical study of existing smart city evaluation frameworks. The paper classifies and compares frameworks according to domains such as governance, technology, sustainability, mobility, economy, quality of life, and social inclusion.	It identifies deficiencies in existing frameworks, including lack of standardisation, social equity, and citizen participation. It highlights the need for a unified and inclusive evaluation framework.	Analytical study, without implementation or validation. It does not propose or test a new operational framework. The results depend on the selection of the literature.
[64]	Smart city governance and interoperability: enhancing human security in Yogyakarta and Makassar, Indonesia	ICT platforms, interoperable information systems, data integration platforms, e-government systems, smart city platforms, PLS-SEM analytical tools.	It analyses how interoperability attributes influence smart city governance and human security. The study is based on survey data collected from 315 respondents across 47 government agencies in Yogyakarta and Makassar, using PLS-SEM modelling.	The results demonstrate that smart city governance enhances human security. Design processes, user interaction, service consumption, and change management show strong positive effects.	The study is limited to two cities in Indonesia. It is focused on the public sector and is limited with regard to the private sector and civil society. It does not evaluate smart city technologies at the operational level.
[65]	Coupled evaluation and forecasting of smart city sustainability using Kolmogorov-Arnold networks	IoT, ICT infrastructure, Big Data, TOPSIS, entropy weighting method, coupling coordination model, Kolmogorov-Arnold Networks (KANs), AI-based time-series forecasting.	Evaluation of cities using TOPSIS, entropy weighting, coupling coordination analysis, and KANs. It assesses Urban Smart Infrastructure (USI), Environmental Pollution Control (EPC), and Urban Ecological Greening (UEG) across 21 smart cities in China.	Improvement in the coordination levels of smart cities, with forecasts indicating growth towards advanced coordination by 2030. KANs outperform MLP, CNN, and LSTM models in terms of forecasting accuracy and convergence.	Limited to smart city pilot regions in China. The results depend on the availability and quality of statistical yearbook data. It focuses on sustainability indicators and does not directly evaluate performance.
[66]	Integration of the Internet of Things in smart cities: enhancing urban life through connected technologies	IoT, AI, ML, edge computing, cloud computing, blockchain, low-power networks, 5G, IoT sensors powered by renewable energy sources, MQTT, CoAP, HTTP.	It proposes a multi-layer, AI-enhanced IoT framework for smart cities that integrates heterogeneous urban systems. The architecture focuses on interoperability, cybersecurity, energy efficiency, and citizen-centred engagement.	The evaluations indicate interoperability across platforms, reduction of security risks by 85%, reduction in energy consumption by 12–40%, improvement in resource efficiency by 35%, reduction in maintenance costs by 50%, and reduction in traffic congestion by 30%.	The results are validated through testing and simulated urban environments rather than real urban deployments. The framework’s performance depends on institutional capacity.
[67]	MACeIP: a context-enriched multimodal ambient intelligence platform for smart cities	IoT sensors, multimodal AI, XAI, edge computing, cloud computing, LoRaWAN, computer vision, LSTM forecasting, AR/VR interfaces, APIs.	Multimodal platform for smart cities that supports urban management. It includes Interactive Hubs, a pedestrian monitoring system, intelligent public asset management, and a City Planning Portal (CPP).	The prototype counts pedestrians in real time and detects abnormal behaviour, forecasts energy demand for street lighting, monitors environmental parameters, and improves interaction through Interactive Hubs.	Prototype implementation, without a pertinent evaluation. There is no comparison with alternative platforms or KPIs. Cybersecurity and privacy are addressed conceptually.

, an analytical synthesis of the most relevant recent contributions dedicated to the implementation of the technologies necessary for an integrated smart city system is presented. The corpus covers architectural frameworks and methodologies, as well as experimental implementations validated by simulated results or by isolated testing.

Table 1 highlights three recurring patterns in recent smart city implementations and testbeds.

First, many contributions emphasise infrastructural capabilities (communications, sensing, data collection) and interoperability enablers yet report limited outcome indicators at city level. KPIs are not strongly represented in the analysed research; instead, there is a tendency towards general analyses related to service adoption, equity impacts, or safety outcomes.

Second, cross-domain integration remains partial. Even when systems are described as “platforms”, the evidence typically covers only a subset of domains, with limited discussion of unified data models and end-to-end workflows encompassing e-government, mobility, healthcare, energy, and public safety.

Third, governance and data stewardship are unevenly addressed. Several studies acknowledge issues related to privacy, security, and organisational barriers, but few operationalise data governance as first-class architectural components alongside sensing, networking, and analytics. Moreover, the analysed works tend to focus on only a small part of what would constitute a complete future city system. As a result, the multiple aspects involved in the design of an integrated smart city system are either insufficiently addressed or only marginally analysed.

2.3. Research Gaps and Derived Requirements for Smart City Platforms

Based on the thematic review and the analysis presented in Table 1, the following gaps are identified within the reviewed research:

-

Lack of city-level KPIs: Many studies do not define a clear evaluation framework beyond technical link metrics. Some research presents multiple data analyses but leaves room for interpretation, which reduces the credibility of the obtained results. In contrast, specifying a minimal set of KPIs such as latency, availability, scalability, robustness, and safety, together with the definition of a monitoring layer for collecting these indicators, would have strengthened the examined approaches;

-

Fragmentation across domains: Implementations are frequently single-domain or weakly integrated. In such a field, the adoption of a more comprehensive architecture is required, enabling modularisation and cross-domain integration, with explicit integration patterns and shared services;

-

Interoperability is mentioned but not operationalised: Standards are cited, yet data harmonisation and semantic alignment are often insufficiently defined. Addressing semantic data layers and well-defined API contracts is necessary to enable seamless exchange between heterogeneous systems;

-

Data governance is insufficiently specified: Privacy, access rights, and auditing are rarely treated as fundamental architectural elements. The integration of data governance and security-by-design (roles, policies, consent, and audit logs) within a multi-layer system must be analysed;

-

Limited discussion of scalability: Many results stem from pilot projects or short-term evaluations without addressing issues such as horizontal scaling, graceful degradation, and operational resilience;

-

Limited linkage to user needs and public policies: Technical components are not always aligned with citizen-centred services and public policy objectives.

In

Table 2. Thematic coding matrix of references from the literature.

No.	Category	References	Observations/Role in Research
1	IoT & networks/infrastructure	[38,39,41,43,45,46,48,66,67]	Technological foundations for smart city infrastructure and data collection and transmission.
2	Platforms/testbeds & living labs	[38,39,45,46,67]	Experimental validation and real-world implementations of smart city solutions.
3	Digital twin/3D modelling/metaverse	[42,49,58,59]	Virtual representation of the city for analysis, simulation, and intelligent management.
4	AI/Edge AI/AIoT/XAI	[40,55,57,58,59,62,65,66,67]	Intelligent decision support, automation, and advanced analysis of urban data.
5	KPI/indicators/evaluation & benchmarking	[52,53,54,60,63,65]	Measurement of smart city performance, sustainability, and maturity.
6	Governance/interoperability/e-government	[41,47,50,54,56,64]	Institutional framework, public policies, and interoperability between systems.
7	Resilience/crisis & emergency management	[51,56,62]	Enhancing urban response capacity to crises and emergency situations.
8	Case studies (cities/countries) & public policies	[44,47,51,62,64]	Practical applicability and lessons learned from real-world implementations.
9	Bibliometric analyses/trends	[61]	Identification of research directions and the evolution of the smart city field.

, the analysed references are grouped according to their domain of application, providing an overview of the main research directions in smart cities.

As presented in this section, the literature indicates significant progress in sensors, connectivity, and domain-specific services, while also highlighting persistent fragmentation, uneven governance practices, and a limited capacity for comparable evaluation at the urban level. These gaps motivate the hub-oriented infrastructural framework proposed in this paper and underpin the methodological choices and design priorities addressed in the following sections.

3. Materials and Methods

3.1. Common Methods Used in Smart City Systems

The design of an urban informational hub is based on the initial processes of cleaning and unifying heterogeneous data sets, followed by the identification of significant deviations and the anticipation of short-term developments. On this basis, policies are defined for urban mobility and energy, while resources are allocated so as to balance performance, costs, and social equity.

As shown in the previous section of the specialised literature regarding smart city approaches and platforms and hub-based or unified system architectures, we can outline an idea of the direction of these systems. At the same time, a more detailed analysis is necessary in order to understand how those systems operate. For this reason, this section addresses two approaches, presented in the two main subsections.

The first approach highlights the methods, techniques, strategies and algorithms most frequently reported as effective for real-time urban monitoring and decision support, arranged in the form of an operational pipeline in accordance with typical data flows. This stage is strictly linked to and consistent with the research conducted in the specialised study.

The second approach addresses the directions for conceptualising the system proposed in this paper, drawing a parallel with what other authors have proposed in other works.

3.1.1. Data Cleaning, Fusion and Quality

Urban sensing data collected from heterogeneous IoT devices are usually noisy, incomplete and asynchronously sampled. The analysed studies consistently report the need for robust cleaning, fusion and qualitative assessment for reliable anomaly detection, prediction and control within informational hubs. The methods below synthesise frequently used techniques for data smoothing, multi-sensor fusion and state estimation.

1. 

Exponential moving average (smoothing/imputation):

Smoothing/imputation of missing sensor values (see Equation (1)) [68,69,70].

$${\hat{x}}_t=\alpha x_t+(1-\alpha ){\hat{x}}_{t-1}$$

(1)

Terms:

• $x_t$: observed value at time $t$ (e.g., $P{M}_{2.5}$ concentration);
• ${\hat{x}}_t$: smoothed/estimated value at $t$;
• $\alpha \in \left(\mathrm{0,1}\right]$: smoothing factor (higher = faster response);
• ${\hat{x}}_{t-1}$: previous estimate.

2. 

Sensor fusion (inverse-variance weighting):

Combining measurements from multiple sensors into a more robust one (see Equation (2)) [71,72].

$$w_i={\displaystyle \frac{{\sigma }_i^{-2}}{\sum _{j=1}^m{\sigma }_j^{-2}}},\hat{x}=\sum _{i=1}^m{w}_i{x}_i$$

(2)

Terms:

• $x_i$: measurement from sensor $i$;
• ${\sigma }_i^2$: error variance of sensor $i$ (lower = better);
• $w_i$: normalised weight (sums to 1);
• $\hat{x}$: fused estimate.

3. 

Kalman filter (linear and discrete)—prediction and update:

Estimating the real state from noisy measurements (filtering and correction) [73,74].

Prediction (see Equation (3)):

$${\hat{\mathbf{x}}}_{t|t-1}=\mathbf{A}{\hat{\mathbf{x}}}_{t-1|t-1}+\mathbf{B}{\mathbf{u}}_t,{\mathbf{P}}_{t|t-1}=\mathbf{A}{\mathbf{P}}_{t-1|t-1}{\mathbf{A}}^{\mathrm{\top }}+\mathbf{Q}$$

(3)

Update (see Equations (4) and (5)):

$${\mathbf{K}}_t={\mathbf{P}}_{t|t-1}{\mathbf{H}}^{\mathrm{\top }}(\mathbf{H}{\mathbf{P}}_{t|t-1}{\mathbf{H}}^{\mathrm{\top }}+\mathbf{R}{)}^{-1}$$

(4)

$${\hat{\mathbf{x}}}_{t|t}={\hat{\mathbf{x}}}_{t|t-1}+{\mathbf{K}}_t({\mathbf{z}}_t-\mathbf{H}{\hat{\mathbf{x}}}_{t|t-1}),{\mathbf{P}}_{t|t}=(\mathbf{I}-{\mathbf{K}}_t\mathbf{H}){\mathbf{P}}_{t|t-1}$$

(5)

Terms:

• ${\hat{\mathbf{x}}}_{t|t-1}$: predicted state;
• ${\hat{\mathbf{x}}}_{t|t}$: corrected state;
• $\mathbf{A}$: state transition;
• $\mathbf{B}$: input matrix;
• ${\mathbf{u}}_t$: control;
• $\mathbf{P}$: state error covariance;
• $\mathbf{Q}$: process noise covariance;
• $\mathbf{H}$: observation matrix;
• ${\mathbf{z}}_t$: measurement;
• $\mathbf{R}$: measurement noise covariance;
• ${\mathbf{K}}_t$: Kalman gain.

4. 

Data quality indicators:

Assessing data quality (completeness, error relative to benchmark) [75,76].

Completeness (see Equation (6)):

$$\mathrm{C}\mathrm{o}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}\mathrm{t}\mathrm{e}\mathrm{n}\mathrm{e}\mathrm{s}\mathrm{s}={\displaystyle \frac{N_{\mathrm{v}\mathrm{a}\mathrm{l}\mathrm{i}\mathrm{d}}}{N_{\mathrm{e}\mathrm{x}\mathrm{p}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{e}\mathrm{d}}}}$$

(6)

Root Mean Square Error (RMSE) (see Equation (7)):

$$\mathrm{R}\mathrm{M}\mathrm{S}\mathrm{E}=\sqrt{{\displaystyle \frac{1}{n}}\sum _{t=1}^n(y_t-{\hat{y}}_t{)}^2}$$

(7)

Terms:

• $N_{\mathrm{v}\mathrm{a}\mathrm{l}\mathrm{i}\mathrm{d}}$: number of valid values received;
• $N_{\mathrm{e}\mathrm{x}\mathrm{p}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{e}\mathrm{d}}$: expected values (from sampling frequency);
• $y_t$: reference “ground-truth” value;
• ${\hat{y}}_t$: measured/estimated value;
• $n$: number of observations.

3.1.2. Anomaly Detection from Environment, Energy and Traffic

After the data pre-processing stage, anomaly detection is used to identify sensor faults and unusual events. In the analysed literature, robust statistics and decomposition-based approaches are frequently adopted because they can operate online and are less sensitive to extreme values. Therefore, these stages must be addressed in advance in the design of a system, since an unsound informational basis will lead to imprecise results.

1. 

Robust Z-score:

Automatic detection of abnormal values (outliers), robust to extremes (see Equation (8)) [77,78].

$$z_t={\displaystyle \frac{x_t-\tilde{x}}{\mathrm{M}\mathrm{A}\mathrm{D}}},\mathrm{M}\mathrm{A}\mathrm{D}=\mathrm{m}\mathrm{e}\mathrm{d}\mathrm{i}\mathrm{a}\mathrm{n}(|x-\tilde{x}|)$$

(8)

Terms:

• $x_t$: value at time $t$;
• $\tilde{x}$: series median;
• $\mathrm{M}\mathrm{A}\mathrm{D}$: median absolute deviation (scale);
• Flag anomaly when $\left|z_t\right|>k$ (e.g., $k=3$).

2. 

Seasonal-Trend decomposition using Loess (STL):

Separating trend and detecting anomalies in the residual (see Equation (9)) [79,80].

$$x_t=T_t+S_t+R_t$$

(9)

Terms:

• $T_t$: trend; $S_t$: seasonal component;
• $R_t$: residual (irregular).

3.1.3. Time Series Forecasting

Short-term predictions support proactive decision-making such as anticipating congestion, energy peaks or the deterioration of air quality. The specialised literature often reports classical statistical models such as ARIMA for univariate series, complemented by clear error metrics to compare forecast accuracy across studies and implementations.

1. 

AutoRegressive Integrated Moving Average (ARIMA)—ARIMA (p,d,q):

Short-term forecasting of univariate time series (see Equation (10)) [81,82].

$$\varphi \left(B\right)(1-B{)}^d{x}_t=\theta (B){\epsilon }_t$$

(10)

Terms:

• $B$: backshift operator ($B{x}_t=x_{t-1}$);
• $d$: differencing order;
• $\varphi \left(B\right)$: AR polynomial;
• $\theta \left(B\right)$: MA polynomial;
• ${\epsilon }_t$: white noise (zero mean, variance ${\sigma }^2$).

2. 

Error metrics RMSE and Mean Absolute Percentage Error (MAPE):

Evaluating the accuracy of forecasts (error magnitude) (see Equation (11)) [83,84].

$$\mathrm{R}\mathrm{M}\mathrm{S}\mathrm{E}=\sqrt{{\displaystyle \frac{1}{n}}\sum _{t=1}^n(y_t-{\hat{y}}_t{)}^2},\mathrm{M}\mathrm{A}\mathrm{P}\mathrm{E}={\displaystyle \frac{100}{n}}\sum _{t=1}^n\left|{\displaystyle \frac{y_t-{\hat{y}}_t}{y_t}}\right|$$

(11)

Terms:

• $y_t$: actual value;
• ${\hat{y}}_t$: forecast;
• $n$: number of points.

3.1.4. Intelligent Mobility

Mobility management is based on routing and control mechanisms that can incorporate multiple objectives such as travel time, emissions generated or system reliability. They can also adapt to demand, being effective in the process of change. The analysed research highlights multi-criteria route planning, predictive control for traffic lights and reinforcement learning as effective approaches for dynamic traffic management.

1. 

Multi-criteria route cost (for Dijkstra/A*):

Computation of the optimal route with multiple objectives (time/emissions/variability) (see Equation (12)) [85,86].

$$c_e=w_t·{\mathrm{t}\mathrm{i}\mathrm{m}\mathrm{e}}_e+w_e·{\mathrm{e}\mathrm{m}\mathrm{i}\mathrm{s}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{s}}_e+w_v·{\mathrm{v}\mathrm{a}\mathrm{r}\mathrm{i}\mathrm{a}\mathrm{b}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{t}\mathrm{y}}_e$$

(12)

Terms:

• $e$: road segment;
• ${\mathrm{t}\mathrm{i}\mathrm{m}\mathrm{e}}_e$: travel time;
• ${\mathrm{e}\mathrm{m}\mathrm{i}\mathrm{s}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{s}}_e$: emissions;
• ${\mathrm{v}\mathrm{a}\mathrm{r}\mathrm{i}\mathrm{a}\mathrm{b}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{t}\mathrm{y}}_e$: uncertainty/variance;
• $w_t,w_e,w_v\ge 0$, typically $\sum w=1$.

2. 

Model Predictive Control (MPC) for traffic signals:

Dynamic traffic signal optimisation (predictive control) (see Equations (13) and (14)) [87,88].

$$\underset{u_{1:T}}{min}\sum _{t=1}^T({\lambda }_q\Vert {\mathbf{q}}_t{\Vert }_2^2+{\lambda }_u\Vert \mathsf{\Delta }{\mathbf{u}}_t{\Vert }_2^2)$$

(13)

subject to

$${\mathbf{q}}_{t+1}=f({\mathbf{q}}_t,{\mathbf{u}}_t,{\mathbf{d}}_t),u_{min}\le {\mathbf{u}}_t\le u_{max}$$

(14)

Terms:

• ${\mathbf{q}}_t$: queue vector;
• ${\mathbf{u}}_t$: control (green/red times);
• $\mathsf{\Delta }{\mathbf{u}}_t={\mathbf{u}}_t-{\mathbf{u}}_{t-1}$: change in control;
• ${\mathbf{d}}_t$: disturbances (demand);
• $f(\cdot )$: network model;
• ${\lambda }_q,{\lambda }_u\ge 0$: weights;
• $T$: horizon, bounds ensure safety.

3. 

Reinforcement learning—Bellman optimality (Q*):

Learning a control policy from rewards (adaptive traffic) (see Equation (15)) [89,90].

$$Q^{*}(s,a)=\mathbb{E}[r+\gamma \underset{a^{\prime }}{max}{Q}^{*}(s^{\prime },a^{\prime })\mid s,a]$$

(15)

Terms:

• $s$: state;
• $a$: action;
• $r$: reward;
• $s^{\prime }$: next state;
• $\gamma \in \left[\mathrm{0,1}\right)$: discount.

3.1.5. Energy and Buildings

Some research focuses on energy management and considers it a point of interest in smart city systems. This type of management concentrates on reducing costs and emissions while maintaining occupants’ comfort. Many of the analysed works formulate energy scheduling and demand response as optimisation problems with constraints derived from the thermal dynamics of buildings and device limitations.

1. 

Optimal control of energy consumption (DR/HP scheduling):

Planning/optimisation of consumption and cost reduction while maintaining comfort (see Equation (16)) [91,92].

$$min\sum _t(p_t{E}_t+\alpha |\mathsf{\Delta }{E}_t|+\beta {\mathrm{d}\mathrm{i}\mathrm{s}\mathrm{c}\mathrm{o}\mathrm{m}\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{t}}_t)$$

(16)

Terms:

• $p_t$: energy price (€/kWh);
• $E_t$: energy at $t$; $\mathsf{\Delta }{E}_t=E_t-E_{t-1}$;
• ${\mathrm{d}\mathrm{i}\mathrm{s}\mathrm{c}\mathrm{o}\mathrm{m}\mathrm{f}\mathrm{o}\mathrm{r}\mathrm{t}}_t$: comfort penalty (e.g., $PMV/PPD$ or $|T_{\mathrm{i}\mathrm{n}\mathrm{t},t}-T^{\star }|$);
• $\alpha ,\beta \ge 0$ weights; subject to comfort/power constraints.

Constraints: maximum power, time windows, thermal dynamics of the building (see Equation (17)):

$$T_{\mathrm{int},t+1}=a{T}_{\mathrm{int},t}+b{T}_{\mathrm{ext},t}+c{P}_t+{\xi }_t$$

(17)

where $T_{\mathrm{int}}(°\mathrm{C})$ is the indoor temperature, $T_{\mathrm{ext}}$ the outdoor temperature, $T_{\mathrm{ext}}$ the HVAC power, and $T_{\mathrm{ext}}$ the noise.

3.1.6. Computer Vision in Road Safety and Parking

Computer vision pipelines are widely used for monitoring road safety, such as the prior detection of accidents, incident analysis and the estimation of parking occupancy using cameras. The literature usually evaluates segmentation-based detection models using standard metrics such as IoU and mean average precision, allowing comparable reporting of performance.

1. 

Intersection over Union (IoU) and detection metrics:

Evaluation of object detection/segmentation (see Equations (18) and (19)) [93,94].

$$\mathrm{I}\mathrm{o}\mathrm{U}={\displaystyle \frac{|B\cap B^{*}|}{|B\cup B^{*}|}}$$

(18)

Terms:

• $B$: predicted box;
• $B^{*}$: ground-truth box;
• $|\cdot |$: area.

$$\mathrm{P}\mathrm{r}\mathrm{e}\mathrm{c}\mathrm{i}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}={\displaystyle \frac{TP}{TP+FP}},\mathrm{R}\mathrm{e}\mathrm{c}\mathrm{a}\mathrm{l}\mathrm{l}={\displaystyle \frac{TP}{TP+FN}}$$

(19)

Terms:

• $TP$: true positives; $FP$: false positives; $FN$: false negatives;
• $AP$: area under Precision-Recall;
• $mAP$: mean $AP$ across classes.

3.1.7. Optimisation of Resources and Services

Urban services such as fleet management, maintenance dispatching and emergency response require the allocation of limited resources in space and time. The analysed studies frequently use multi-objective optimisation, fairness metrics and combinatorial allocation or routing formulations in order to balance efficiency, cost and equity. All these methods can be gradually integrated into an intelligent management system as long as the level of implementation allows modularisation.

1. 

Aggregation of multiple objectives (weighted-sum):

Aggregation of multiple objectives into one optimised criterion (see Equation (20)) [95,96].

$$\underset{x}{min}\sum _{k=1}^K{w}_k{f}_k\left(x\right) \mathrm{with} w_k\ge 0,\sum w_k=1$$

(20)

Terms:

• $x$: decision vector;
• $f_k\left(x\right)$: objective $k$ (e.g., time, emissions, cost);
• $w_k$: weight of objective $k$.

2. 

Fairness (Jain’s index):

Measuring the fairness of resource allocation among entities (see Equation (21)) [97,98].

$$J={\displaystyle \frac{(\sum _{i=1}^n{x}_i{)}^2}{n\sum _{i=1}^n{x}_i^2}},J\in (\mathrm{0,1}]$$

(21)

Terms:

• $x_i$: allocation to entity $i$;
• $n$: number of entities;
• $J\to 1$ means fair.

3. 

Task assignment:

Optimal allocation of agents–tasks given a cost (see Equations (22) and (23)) [99,100].

$$\underset{x_{ij}}{min}\sum _{i=1}^N\sum _{j=1}^M{c}_{ij}{x}_{ij}$$

(22)

subject to

$$\sum _j{x}_{ij}=1,\sum _i{x}_{ij}=1,x_{ij}\in \left\{\mathrm{0,1}\right\}$$

(23)

Terms:

• $i$: agents/teams, $j$: tasks/locations;
• $c_{ij}$: cost of assigning agent $i$ to task $j$ (time/distance);
• $x_{ij}$: binary decision (1 if agent $i$ takes task $j$).

4. 

Vehicle Routing (capacity constraint, simplified):

Efficient fleet routing with capacity constraints (see Equations (24) and (25)) [101,102].

$$\underset{y_{ij}}{min}\sum _i\sum _j{d}_{ij}{y}_{ij}$$

(24)

subject to

$$\sum _j{y}_{ij}=1,\sum _i{y}_{ij}=1,\sum _{i \in \mathrm{r}\mathrm{o}\mathrm{u}\mathrm{t}\mathrm{e}}{\mathrm{d}\mathrm{e}\mathrm{m}}_i\le Q,y_{ij}\in \left\{\mathrm{0,1}\right\}$$

(25)

Terms:

• $d_{ij}$: distance/cost from $i$ to $j$;
• $y_{ij}$: chosen arc (0/1);
• ${\mathrm{d}\mathrm{e}\mathrm{m}}_i$: demand at point $i$;
• $Q$: vehicle capacity.

3.2. Candidate Methods for the Conceptual Smart City Hub Prototype

The conceptual prototype includes various applicable methods that serve modular development at system level. The methods are presented in basic flows, and the data originating from sensors and the integration of maps highlight the spatial dimension.

Based on the analysis of the specialised literature in the dedicated section and the methods extracted from the studies presented in the previous subsection, this stage will refer to the approaches chosen in outlining, at the level of a conceptual prototype, the proposed informational hub-type system that achieves the gradual transition from an ordinary city to a smart city.

The proposed system integrates representative services such as platform authentication, vehicle parking, the purchase of tickets for public transport and payments related to the purchase of services. Likewise, the main analytical sensing layers are related to the environment, geospatial optimisation and event notifications.

At this stage, the prototype is conceived as a reference concept and a concise description at the level of methods used; however, the subject is not focused on the actual implementation and does not provide quantitative performance results, but it offers a solid informational basis for designing this type of system.

3.2.1. Authentication and Security

Authentication and security are cross-cutting requirements in intelligent platforms, and the analysed implementations usually place them in a dedicated identity and access service in order to support interoperability between applications or subsystems.

The following equation indicates that the stored value is a one-way digest of the password, combined with a unique salt per user. This prevents the storage of passwords in plain text and reduces the impact of rainbow-table attacks. In a hub, as well as in this proposed system, it is a mechanism that would be part of the authentication service together with secure password policies and rate limiting, as recommended in the analysed security-focused studies.

For storing and verifying user passwords, cryptographic hash functions are applied (see Equation (26)) [103,104]:

$$H\left(p\right)=SHA256(p\parallel salt)$$

(26)

where:

• $H\left(p\right)$: resulting hash for the user’s password;
• $p$: user’s input password;
• $salt$: a random value unique per user;
• $\parallel$: concatenation operator.

3.2.2. Parking Management

Vehicle parking management is also another service that should be part of a smart city of the future, as it is frequently oriented towards citizens.

The following equation models parking charges as a base fee plus a component proportional to time. This is a common pricing mechanism in parking management systems and can be integrated with occupancy sensing flows and the enforcement of service usage rules.

The parking cost is calculated according to the duration of use (see Equation (27)) [105,106]:

$$C=T_{base}+R\times \mathsf{\Delta }t$$

(27)

where:

• $C$: total parking cost;
• $T_{base}$: base (minimum) fee;
• $R$: tariff per unit of time (per hour or minutes);
• $\mathsf{\Delta }t$: parking duration (in hours or minutes).

Dynamic tariffs can be modelled as a piecewise function (see Equation (28)):

$$C(\mathsf{\Delta }t)=\left\{\begin{array}{cc}{R}_1\times \mathsf{\Delta }t,& \mathsf{\Delta }t\le 2 \mathrm{h}\\ R_2\times \mathsf{\Delta }t,& \mathsf{\Delta }t>2 \mathrm{h}\end{array}\right.$$

(28)

where $R_1$ and $R_2$ are different rates depending on the duration.

3.2.3. Transport and Electronic Tickets

Digital ticketing for public transport is a useful service for citizens as it offers faster accessibility. This service is commonly based on unique, verifiable identifiers that can be issued and validated quickly, including in offline scenarios or when connectivity is intermittent.

The presented equation generates a ticket identifier by hashing the user ID, the time and a random nonce. This supports the traceability and uniqueness of the identifiers used, and the approach can be easily integrated as a Quick Response (QR) code or Near Field Communication (NFC), reducing the risk of forgery.

A unique ticket is generated through a hash function (see Equation (29)) [107,108]:

$$TicketID=H(userId\parallel timestamp\parallel random)$$

(29)

where:

• $TicketID$: unique identifier of the ticket;
• $userId$: user’s ID;
• $timestamp$: generation time;
• $random$: random value ensuring uniqueness.

Validation is defined as (see Equation (30)) [109,110]

$$valid=(now\le validUntil)\wedge checkHash\left(TicketID\right)$$

(30)

where:

• $now$: current time;
• $validUntil$: expiration time of the ticket;
• $checkHash\left(TicketID\right)$: function verifying the ticket’s integrity;
• $\wedge$: logical AND operator.

3.2.4. Payment Processing

Payment processing is usually managed as a transactional flow with strong consistency and audit requirements, and modelling the payment life cycle helps to avoid duplicate charges and supports reliable retries. This service must also be addressed, as the citizen must have the freedom to pay bills in a centralised, rapid, secure and efficient manner. Financial transactions can be represented by the equation below, which models payment processing as controlled transitions between well-defined states. This reduces ambiguity in distributed systems and supports auditability and idempotent processing, which are common design recommendations in payment systems engineering.

Financial transactions are represented as a Finite State Machine (FSM) (see Equation (31)) [111,112]:

$$\{pending\to authorised\to settled\}$$

(31)

where:

• $pending$: payment initiated but not yet confirmed;
• $authorised$: payment authorised by the processor;
• $settled$: payment completed and confirmed.

3.2.5. Sensors and Environmental Data

The sensing layer aggregates heterogeneous streams (environment, mobility, infrastructure) and prepares them for the analytical process. Light data smoothing and basic anomaly flags can be used as subsequent checks before advanced modelling. Moving averages provide a simple temporal aggregation to smooth high-frequency sensor noise and to produce stable indicators.

This is frequently used in intelligent system dashboards and as a pre-processing stage for models developed in other stages.

Sensor values are aggregated using moving averages (see Equation (32)) [113,114]:

$$M{A}_t={\displaystyle \frac{1}{n}}\sum _{i=0}^{n-1}{x}_{t-i}$$

(32)

where:

• $M{A}_t$: moving average at time $t$;
• $n$: size of the time window;
• $x_{t-i}$: sensor value at step $t-i$.

For anomaly detection, standard deviation is applied (see Equation (33)) [115,116]:

$$\sigma =\sqrt{{\displaystyle \frac{1}{N}}\sum _{i=1}^N(x_i-\mu {)}^2}$$

(33)

where:

• $\sigma$: standard deviation;
• $N$: number of observations;
• $x_i$: value of each observation;
• $\mu$: mean of the values.

A sensor reading is considered anomalous if (see Equation (34)):

$$|x_t-\mu |>k\cdot \sigma$$

(34)

where $k$ is a sensitivity factor (e.g., 2 or 3).

3.2.6. Geospatial Optimisation

Geospatial optimisation supports location-based decisions and is usually implemented with graph algorithms and spatial indices in order to ensure real-time response at city scale.

The presented equation captures a distance-minimisation objective for allocating vehicle parking spaces to users. In a complete hub, this formulation can be extended with constraints such as capacity, priority users, reservation rules and network distances. However, geospatial optimisation can be used not only in this sector, as it can be easily integrated into several fields of activity that require allocation, positioning and the identification of the appropriate location.

The allocation of vehicles to parking spots is expressed as a minimisation problem (see Equation (35)) [117,118]:

$$min\sum _{i=1}^{n}d(u_i,p_j)$$

(35)

where:

• $u_i$: position of user $i$;
• $p_j$: position of parking spot $j$;
• $d(u_i,p_j)$: distance between user and spot;
• $n$: total number of users.

Dijkstra’s algorithm is applied for shortest path computation, and R-tree structures for spatial queries.

3.2.7. Notifications and Events

In the proposed system, it is necessary to achieve real-time communication of events in order to optimise processes. Event-based communication is widely used to decouple services and to enable scalability, fault isolation and notifications to users.

In the presented equation, event delivery is formalised in an architecture based on data transmission. A service receives an event if it is subscribed, and the message is recorded as delivered up to time t. This reflects common message bus patterns used to decouple services (parking, payment, notifications) and to support scalability, as discussed in the literature on Smart City platform architectures.

Events such as ParkingStarted or PaymentSettled are propagated through a message bus. The delivery condition is (see Equation (36)) [119,120]

$$delivered\left(e\right)⟺\forall s\in Subscribers,\exists t:e\in Lo{g}_s\left(t\right)$$

(36)

where:

• $delivered\left(e\right)$: event $e$ has been delivered;
• $Subscribers$: set of subscribed services;
• $Lo{g}_s\left(t\right)$: log of messages received by service s until time $t$.

3.3. Approaches Strategies

The approach strategies for the development and validation of an informational and operational hub-type system are based on the principle of gradual evolution, highlighting the working stages in a detailed manner as suggested in the analysed literature regarding large-scale urban systems.

This paper presents the idea of developing such a system that encompasses subsystems from different areas of interest and consolidates them in order to aggregate data into an operational centre dedicated to smart city administration. The paper is oriented towards a proposal rather than an actual implementation and highlights the development steps for the operationalisation of a hub-type system.

The system components can be implemented incrementally as coupled services and connected through an event-driven architecture, while at the same time integrating external systems and interoperable data models. This type of system must allow the verification of critical flows and the coherence of the user interface. Moreover, authentication and security protocols must be designed so as to ensure data confidentiality, communication integrity and service availability without affecting the user experience.

Another essential strategy is the adoption of an event-driven architecture, which allows asynchronous management of actions and the scaling of components without affecting overall performance. In parallel, the use of geospatial optimisation methods and the analysis of data collected from sensors support real-time decision-making. By combining these strategies, the system can evolve from a demonstrative implementation towards a robust, scalable platform that can be integrated into complex urban ecosystems.

As presented in this section, the preliminary methods, techniques and strategies for designing a system represent the essential building blocks in constructing a coherent and scalable architecture capable of supporting the gradual integration of urban services and interoperability between subsystems.

4. The Proposed Smart City Hub System and Its Conceptual Architecture

4.1. General Description of the Architecture

The proposed system, called Smart City Hub, is at an early stage and at concept level, serving as an integrated digital platform intended to centralise, coordinate and simplify the interaction between citizens, municipal services and smart infrastructures. Its purpose is to provide a single point of access to urban functionalities, contributing to increased efficiency, transparency and citizen engagement.

The platform is conceived as a responsive web application, accessible from a wide range of devices such as desktop, tablet and smartphone. It also ensures usability in different contexts. Its design is based on the principles of inclusion and accessibility, allowing citizens with varying levels of digital literacy to easily interact with municipal services.

At its core, the system supports several functional areas representative of a smart urban ecosystem. These include city hall services (notices, complaints, appointments), transport (ticket purchase and validation, mobility integration), payments (utilities, local taxes), parking management, map services, Smart Home integration, community engagement and personalized notifications. Together, these modules cover the main levels of interaction between citizens and the urban environment.

Smart City Hub is not a static platform but an extensible system. Its modularity allows for the gradual addition of new services, such as environmental monitoring, blockchain-based verification or integration with urban IoT sensors. This approach ensures long-term scalability and adaptability to the continuous evolution of technologies.

From the user’s perspective, the platform offers a clear and intuitive navigation model, grouping services into logical and easy-to-navigate sections. Each functional area is designed to meet specific needs, while maintaining coherence with the wider ecosystem. The general principle behind this concept is simplicity combined with comprehensiveness: users should be able to perform both routine tasks and more complex actions without significant effort or specialized knowledge.

This conceptual proposal does not serve only as a digital tool for service delivery. Being a facilitator of citizen participation and engagement, by consolidating requests, payments and notifications into a single interface, it increases trust in municipal processes, reduces operational bottlenecks and supports a data-driven urban model.

4.2. Conceptual Architecture

The conceptual architecture is layered and describes the platform functionalities and provides information about the logical interconnection of the layers. Thus, it logically explains the use without fixing the execution technologies or the runtime topology.

The model aims for a clear separation into multiple working layers, as presented:

• Stakeholders and access channels: The main public actor is the citizen who has access to the responsive web interface for devices. The second participant is the municipal official or dedicated system support who has access to back-office views for workflows and approvals. In parallel, a public read-only Application Programming Interface (API) is used, which exposes open data sets for urban indicators. This layer establishes the entry mode into the system, roles and visibility levels.
• Presentation layer (technology agnostic): The interface is divided into independent User Interface (UI) modules: navigation shell, city services (Council), transport/ticketing, payments, parking/smart-home, urban map, account/authentication and notification centre. Each UI module must map 1-to-1 onto an application capability, allowing vertical delivery and testing of each domain. The UI shell manages navigation, visual consistency, and role-based access policies.
• Application layer: Functional domains are modelled as logical services: identity and access (authentication, roles, sessions); requests and workflows (Council); transportation and ticketing; payments and billing; parking (spots, sessions, rates); Smart Home orchestration (device proxies, rules); maps and geospatial (geocoding, layers); notifications (email/SMS/push); and files and documents (uploads, references). The cross-relationships are explicit because parking and transportation require payments. Requests added to the system are linked to uploaded documents. Payments and statuses send notifications to the user account. These dependencies are logical and can be implemented either through direct calls or through events in the final architecture.
• Data layer (logical models): Data is grouped into four domains: operational (ODS) deals with citizens, sessions, requests, and payments; geospatial repository (GEO) deals with parking spaces, zones, and layers; time-series (TS) deals with sensors, telemetry, and vehicles; and document and evidence storage (DOCS). Each capability declares which data domain it consumes/writes, in order to facilitate migration to specialised databases.
• Initial and future integration: The minimum integration stage includes a payment processor (PSP), map/tile provider (OSM), IoT/MQTT broker for devices, municipal registries for verifications, and an open data portal powered by the public API. Optionally, blockchain anchoring of evidence/vouchers (hashes) is foreseen for integrity and non-repudiation, without introducing critical dependencies in operational flows.
• Cross-cutting policies and constraints: Access is based on roles (citizen, official, administrator) and data segments. Notifications are a cross-cutting mechanism, triggered by state changes, such as “ParkingStarted” and “PaymentSettled”. Observability, orchestration, scaling and data governance are not required at this level.
• Conceptual scope delineation: This model establishes responsibilities, interfaces and dependencies at a logical level, so that each functionality can evolve independently. The approach can use microservices or a modular monolith, as long as the contracts between capabilities and the separation of data domains are respected. The guiding principle is high domain cohesion and low coupling between domains.

This preliminary architecture provides the conceptual skeleton of the platform and aligns the actors, interface, capabilities, data and integration, creating the basis for the detailed design and technological decisions in the enterprise architecture. The incipient conceptual architecture can be analysed in more detail by viewing the graphic representation below (see Figure 2). The illustrations were created using diagrams.net, version 29.6.1.

It illustrates the relationships between actors, interfaces, application capabilities, data sources and external integrations, providing a clear overview of the proposed system structure. The diagram supports the theoretical explanations by highlighting the information flows and how the different components cooperate in a unified and coherent framework. It is important to mention that this conceptual architecture represents the initial stage of the ideas that we wish to address in this system. But it does not represent a final architecture, as this paper is oriented towards presenting an architectural overview of a prototype concept without indicating a direct implementation, since it requires testing in real environments, not only simulations.

4.3. System Usage Diagrams

4.3.1. Mode of Use/User Journey

The user opens the system on the Home page, where he finds cards for all the platform modules (Dashboard, Map, Council, Smart Home, Transport, Payments, Community, Notifications, Account/Auth and Tests). Navigation is constant because, in the header, either the avatar with the “Sign out” option is displayed (if authenticated), or the “Sign in” button (if not connected). And for a more illustrative navigation, the side menu can be accessed, which provides immediate access to any of the available sections.

Dashboard offers a synthetic view of requests to the administration, unpaid bills, Smart Home status, mobility and environmental indicators, active alerts and parking situation.

The Map module allows for punctual reporting to coordinates, initiating and stopping a parking session, as well as viewing sensors, alerts and vehicles.

In Council, the user can submit a request, a notification, a document or an appointment and track the processing status.

Smart Home offers commands for turning lights on and off, setting the temperature, and arming or disarming the security system.

In Transport, the user can buy a ticket for a bus, tram, metro, or taxi. They can validate the ticket via QR and consult the transaction history.

Payments allows marking invoices as paid or unpaid and displays due dates and totals.

In Community, the user browses current or upcoming local events, with title, date, and location and can quickly decide whether to participate or look for other information on the city’s public channels.

In Notifications, the user finds all the alerts and messages of the application (payment confirmations, ticket validations, request statuses, parking events). Each message can be marked as “read”, and the unread notification counter in the header is updated immediately.

In Account/Auth, the operations of registration, authentication, password reset, and recovery are found.

And Tests is available only to the developer and runs automated checks for critical functionalities.

From any module, the user can return to Home or navigate to another section using the side menu and persistent header (see Figure 3).

4.3.2. The Proposed Architecture and Technologies

The proposed architecture follows a clear separation of responsibilities and operates on cloud-native principles to support scaling, resilience and extensibility.

The application is delivered in the browser, as a Next.js/React interface written in TypeScript and styled with Tailwind. The system is capable of Client-Side Rendering (CSR)/Server-Side Rendering (SSR) and real-time updates via WebSocket/Server-Sent Events (SSE) clients. User traffic is terminated at the edge through an Edge layer that combines Content Delivery Network (CDN), Web Application Firewall (WAF) and rate-limiting, which reduces latency, filters malicious traffic and protects the back-end from abuse, before redirecting requests to the main infrastructure.

At the heart of the platform is a Kubernetes cluster that exposes a single sign-on via API Gateway/Ingress, enforcing HyperText Transfer Protocol Secure (HTTPS) and managing JSON Web Token (JWT)/OpenID Connect (OIDC)-based authentication.

On top of the services runs a Back-end for Front-end (BFF) layer, accessible via GraphQL or Representational State Transfer (REST), which aggregates responses from multiple microservices into a web-optimized payload and caches responses in Remote Dictionary Server (Redis) for high-traffic routes or semi-static data. This model reduces the number of round-trips in the UI and isolates the front-end from the internal details of the domains.

Functionality is divided into independent microservices, implemented in Node.js or Go, each with clearly delimited responsibilities such as:

• The Authentication service handles OIDC/Open Authorization 2 (OAuth2) and issues JWTs.
• The Council service orchestrates request submission, document attachment, and workflows.
• The Smart Home service acts as a device proxy and rules engine for orders and statuses.
• The Transportation service handles tickets and validation.
• The Payments service handles invoices, transactions, and reconciliation.
• The Parking service manages parking spaces, rates, and sessions.
• The Notifications service distributes push/email/Short Message Service (SMS) messages.
• The Files service handles document uploads and metadata.
• The Map & Geo service provides geocoding, map proxy, and thematic layers.

BFF routes requests to these services and composes coherent responses for UI pages in the user journey.

Data persists across each layer that is specialized in the domain of work. Transactional and identity information is stored in PostgreSQL, where Atomicity, Consistency, Isolation, Durability (ACID) requirements are prioritized.

Telemetry and time streams (sensors, vehicles, validations) are written to Time-scaleDB/InfluxDB for aggregations on time windows and efficient queries. Files and exports are stored in Simple Storage Service (S3)/MinIO Object Storage (MinIO) compatible object storage, and access is achieved through secure references managed by the Files service. A dedicated Redis serves as a data and session cache at the BFF or read-intensive service level. OpenSearch/Elasticsearch is used for full-text search, auditing, and aggregation of application logs, which allows both fast diagnostics and compliance queries.

Inter-domain communication uses an event bus (Kafka/RabbitMQ). Services emit business events at key moments such as creating a request, starting/stopping a parking session, issuing an invoice, or purchasing a ticket. These events trigger asynchronous actions in other services, such as sending notifications, initiating payments, or indexing in the search engine. This decoupling allows independent scaling of components and maintains low latency in critical flows in the interface.

Integration with external systems is handled by dedicated adapters. The payment processor is invoked by the Payments service, with secure callbacks for confirmations and reconciliation.

Email/SMS sending uses external providers, and mobile notifications rely on the Fire-base Cloud Messaging (FCM)/Apple Push Notification service (APN). Maps and tile images are served through a proxy to OpenStreetMap or compatible providers to control caching and usage policies. IoT devices and gateways communicate via a Message Queuing Telemetry Transport (MQTT) broker, to which the Smart Home service subscribes or publishes commands in a controlled manner. Optionally, the platform can anchor technical evidence (e.g., ticket receipts or parking session closures) in a smart contract on an Ethereum Virtual Machine (EVM) chain (Ethereum/Polygon L2) for traceability and non-repudiation.

Observability is built into the platform. Each service exposes data through Prometheus and OpenTelemetry traces, allowing centralised collection in Prometheus and visualisation in Grafana, and distributed traces are correlated in Jaeger to quickly diagnose causes of latencies or errors. Application and infrastructure logs are aggregated in OpenSearch/Elasticsearch, with retention and indexing policies tailored to audit requirements.

The entire design is secure and scalable through “zero-trust” principles on every “hop”, end-to-end Transport Layer Security (TLS) and tokens with controlled duration.

Services are stateless to the extent possible and scale horizontally. Data volumes are separated by classes (transactional, time-series, files), and caches and event-driven architecture reduce the load on synchronous paths.

Scaling is done automatically by the Horizontal Pod Autoscaler (HPA), and resilience is ensured through multi-instance replication, readiness/liveness probes and retry/backoff on inter-service calls. Continuous Integration (CI)/Continuous Delivery (CD) pipelines deliver signed containerized images and run and apply progressive delivery (canary/blue-green) to limit risk.

This diagram is visually described in the figure below (see Figure 4).

4.3.3. End-to-End Flow for Start Parking and Pay

The diagram below shows, step by step, the exchanges between the user, the Next.js interface, the gateway, the BFF, the domain services and the infrastructure until the payment is completed and the parking session is closed.

The process starts with the authentication stage. The user enters the email and password in the interface, and the application sends the request POST/auth/login to the API Gateway. The Gateway requests validation from the Auth Service that reads the profile from PostgreSQL. If the verified data is correct, then OIDC/JWT tokens are issued. The tokens are returned to the client, and the interface opens a WebSocket/SSE channel initiated with JWT for real-time notifications.

The parking starts when the user presses “Start” on an empty spot. The interface sends POST/parking/start {spotId} to the Gateway, which redirects the call to the BFF. The BFF invokes the Parking Service with startSession(spotId, userId). The service creates an active ParkingSession record in PostgreSQL, marks the ParkingSpot as busy and writes a cache entry for the active session to Redis. In parallel, it emits a ParkingStarted event in the Event Bus. BFF receives the session details and sends them to the interface for rendering, while Notifications consumes ParkingStarted from the bus and pushes a message to the client via the WebSocket channel called “Parking started”.

The payment is made at the user’s command through the session called “Pay”. The interface sends POST/payments/charge {sessionId} to the Gateway, then to BFF, which calls the Payments Service with createPayment(sessionId). The service inserts a payment in the “pending” state into PostgreSQL and publishes PaymentCreated to the Event Bus. It then contacts the payment processor (Stripe/Adyen) for authorization and capture. The result is set to “approved”, and the payment is updated to paid = true, PaymentSettled is emitted, and the event is consumed by the Parking Service to mark the session as invoiced and close it (active = false). Notifications emit the receipt via WebSocket to the client and displays “Payment successful”.

Optionally, a hash of the receipt can be anchored on the blockchain. Payments sends the hash to the blockchain component, receives back the tx hash and the on-chain receipt URL, and persists the reference in the database for auditing.

The flow ends in the interface with the receipt display and the current session closing. Throughout this, the Event Bus decouples producers from consumers, while PostgreSQL maintains transactional states and Redis accelerates session reads. And the WebSocket channel provides immediate feedback to the user (see Figure 5).

4.4. Key Observations

The proposed platform clearly separates the capabilities of the prototype system into layers. This reduces coupling and allows independent evolution of the interface, application capabilities and data models.

The BFF model for the web focuses data aggregation and adaptation for the UI, shortening latency and reducing the number of calls between the browser and microservices.

The microservices model delimits responsibilities by domains (authentication, requests to the council, transport, payments, parking, notifications, files, map and geo), so that technical support can develop, test and scale independently and defects remain isolated to the boundaries of the affected service.

The single gateway enforces security policies through TLS, OIDC and JWT. And the WAF enforces traffic limitation, protection from abuse and volumetric attacks.

Authentication and authorization are centralized, and the roles implemented apply granular access control at the endpoint level as well as the object level on the domain.

The event-driven architecture decouples producers from consumers and allows real-time reactions to state changes, especially for parking, payments, and notifications.

Data is stored in categories: operational, geospatial, temporal, and documents. This results in lower costs and better performance for each category.

Redis is used for volatile data and active sessions, reducing the pressure on the transactional base and accelerating interactive flows.

Integration with payment processors is isolated in a dedicated service, so that replacing the provider or adding a new one does not affect the rest of the prototypical system at the incipient level.

Anchoring receipts in the distributed ledger provides credible negation of changes but does not condition the basic operation and can be enabled only where audit requirements justify it.

Observability is visual, searchable logs, and distributed traceability. This reduces the mean time to detect and repair incidents.

Containerization and orchestration allow horizontal scaling across services with different traffic profiles, with costs controlled through auto-scaling based on data.

The UI is responsive and accessibility-oriented, facilitating expansion to other language communities.

Publishing read APIs to the open data portal supports transparency, reuse, and local innovation initiatives.

Data governance addresses collection minimization, limited retention, and pseudonymization for General Data Protection Regulation (GDPR) compliance and privacy protection.

The testing strategy combines contract testing and flow integration testing, which prevents regressions in the chain. It is important to mention that the tests are only at a very early stage, without having relevant quantitative results, because the current stage of the system is conceptual and prototypical, not being tested in any real environment, and the simulations are only limited in many stages.

The main risks are dependency on third-party integrations, data quality from legacy municipal sources, and operational complexity specific to a diverse set of capabilities. But the proposed architecture mitigates them through service isolation, message queues, and continuous monitoring.

5. Conceptual Prototype of the Proposed Smart City Hub System

5.1. General Description of the Prototype

The proposed prototype of the system was designed as a responsive web application, developed on the Next.js/React framework, with an interface and a modular structure for each functional area of the city. The interface integrates reusable components and a common navigation shell, so that the user has rapid and consistent access to all modules: Dashboard, Interactive Map, Council Services, Smart Home, Transport, Payments, Community, Notifications and Account/Authentication. Access to the system is based on an account with distinct roles for the regular user and the operator. And the sessions are protected with tokens that expire in a controlled time. Data is stored in separate models for operational mode, geospatial mode, time series mode and documents. This allows for low-latency page loading and consistent reports.

Functional flows cover requests to the city hall, maps and points of interest, parking with start and stop, payments with receipt, real-time notifications and account management. The conceptual prototype proposes a test payment processor, a map provider based on OpenStreetMap and an event queue that propagates relevant states to the front-end. Notification mechanisms send confirmations and alerts to the application, and critical actions such as initiating a parking session or marking an invoice as paid are logged for auditing. The map allows for visualization of geospatial layers and position validations, and the payment page aggregates balances, due dates, and transaction history per user.

The back-end proposes exposing a single entry point through a gateway that publishes internal APIs and a public read-only API necessary for the open data portal. The BFF service combines data from multiple microservices to deliver compact views in the UI and reduces the number of browser calls. For performance, it is recommended that this conceptual prototype use caches for active sessions and results that can be reused in the short term, and that intensive operations such as parking updates be transmitted through WebSocket or SSE channels.

Sensitive data is minimized, and the prototype applies retention and pseudonymization rules for compliance with data protection regulations. This platform is limited in terms of testing scenarios because it is intended to serve as a proof of concept, and testing would mean full implementation. The intentional limitations of the prototype in the case of full implementation concern industrial scaling, integration with real municipal registers and the activation of anchors in distributed registers.

The purpose of the prototype is to offer an interactive vision of the Smart City Hub concept, presenting both the user experience and the technical feasibility of the microservices architecture and of integration with municipal and third-party systems.

5.2. Implementation of the Demonstrative Conceptual Framework

This system is designed as a Next.js 14 web application on React 18, written in TypeScript and styled with TailwindCSS, with mixed SSR/CSR rendering and a single navigation shell that loads functional modules. The structure of the demonstrative project is minimal and oriented towards the App Router. The app/layout.tsx file sets the metadata and integrates the global styles defined in app/globals.css, and app/page.tsx instantiates the top-level component components/App.tsx that contains all the UI logic. The Next configuration includes routes declared in tsconfig.json, and the CSS pipeline is provided by postcss.config.js and tailwind.config.ts.

The App.tsx component acts as the UI shell and orchestrates the rendering of all the capabilities defined in the conceptual architecture. Navigation is controlled by a page state and displays distinct modules: “Council/Services” for requests and documents; “Smart Home” for lighting, temperature and system arming; “Transport” for ticket purchase and validation; “Payments” for marking invoices as paid or unpaid; “Community” for events; “Map” for geospatial interaction and parking; “Notifications” for real-time messaging; and “Account/Auth” for registration, authentication and reset.

Each module is a separate React function in the same file for simplicity, but their interfaces are strongly typed by the domain models CityRequest, Payment, SmartHomeState, Ticket, CitySensor, ParkingSpot, CityAlert, AppNotification, Vehicle and ParkingSession. This in theory helps the static validation of properties and data flows.

Authentication provides registration, login and password reset with basic validations, with the functions authRegister, authLogin, authForgot and authReset that maintain a serialized index of users under the keys sc_users and sc_current_user. Passwords are stored with a hash and a plain password field only for automated testing.

Notification flows use an in-memory queue and display messages in the “Notifications” panel, mimicking WebSocket/SSE pushes. The Write Event (WE) connection is represented as an effect on the state, and the change of statuses triggers the creation of an AppNotification object.

The Map module has a Lite mode based on OSM iframes for simplicity and a Real mode with Leaflet and react-leaflet lazy loaded if USE_REAL_MAP is set, which allows for layers for parking, sensors, alerts and vehicles. The reportHere, startParking and stopParking functions handle reporting, occupancy and session closure with appropriate notifications.

The Payments module maintains a table of invoices and changes the paid status with notification and update on the Dashboard, and the Transport module generates SC-{random} tickets and validates them with a validUntil field. Smart Home allows control for lighting, temperature and security, and Council/Services sends requests and documents with timestamps and coordinates.

The Dashboard aggregates data from all modules and calculates averages for Particulate Matter 2.5 (PM2.5), and the status of parking, and execution is with setInterval.

The interface uses Card, Badge and SectionTitle with Tailwind responsive layout and persistent Header with avatar and authentication actions. Form validations and time formatting are unified with fmtDate, fmtTime and fmtDateTime set.

5.3. User Interface System Capabilities

5.3.1. Main Interface System

The main interface system represented by the Home page is the entry point to the platform and offers the user quick access to all available services. On the left is the side navigation menu with all sections of the application, and in the central area interactive cards for the main functionalities are displayed. The user can submit requests to the local authorities, access Smart Home control functions, buy transport tickets, pay bills, view the parking map, consult community events and access notifications. The page includes a user authentication confirmation section, displaying the e-mail address (see Figure 6).

5.3.2. Navigation Within the System Pages

Main uses of the user interface (see Figure 7):

• Dashboard: Summary of the user’s status (administrative requests, bills, home status, mobility, environment and parking);
• Map: Interactive geospatial map (coordinate reporting, parking session management, visualisation of spaces and sensors);
• Council: Administrative interface (submission of requests/complaints, document attachment, monitoring request status);
• Smart Home: Control of home devices (lighting, temperature, security);
• Transportation: Digital ticketing (ticket purchase and validation with unique identifier/QR);
• Payments: Bill management (viewing obligations, marking payments, transaction aggregation);
• Community: Civic information (list of local events with essential data);
• Notifications: Alert system (real-time notifications regarding payments, tickets, requests and parking).

5.3.3. Management of the User Account in the Proposed Conceptual System

The interface provides basic functionalities for account management represented by the Account page. Also, within this section, several pages are available to facilitate user needs. Functionalities for user registration, authentication, password reset and password recovery by sending a reset link are included. The interface is structured intuitively, allowing users to manage their access data in a versatile, user-friendly way and to personalise their experience on the platform. This constitutes the foundation of the system’s security and authentication (see Figure 8).

The system described in this section constitutes an implementation-oriented proposal, currently at the stage of a conceptual prototype, without experimental results or quantitative evaluations that would allow validation in real environments or in representative operational contexts. The approach aims to illustrate the methodological, architectural and functional directions of an urban informational hub, without claiming empirically validated performance.

Nevertheless, the proposed conceptual framework highlights the potential of such a system to improve the accessibility of urban services, reduce informational fragmentation and operational bottlenecks, and facilitate interoperability between administration, intelligent infrastructures and citizens. The implementation of such an integrated hub could contribute to simplifying administrative procedures, increasing institutional transparency and expanding citizens’ autonomy in their interaction with public services, reducing the constraints associated with local bureaucracy and supporting the transition towards intelligent, user-centred urban ecosystems.

6. Conclusions and Future Directions for Research and Development

Based on the specialised literature studied and on the analysis carried out in this paper, we can say that next-generation smart cities can no longer be seen as simple technological experiments but as complex models of urban development, in which digital infrastructure, sustainability and citizen participation converge to build resilient and equitable communities.

In the current context of smart cities, technologies such as AI, IoT, blockchain and digital twins enable the transition from reactive models to proactive ones. What is considered improbable to achieve today may tomorrow become an entirely everyday matter, and advanced technologies, used responsibly, can raise the standard of living of citizens exponentially.

These emerging technologies, innovative governance models and interdisciplinary collaborations outline promising directions for the future. The opportunities are not only about increasing the efficiency of infrastructures but also about strengthening social inclusion, environmental protection and innovation-driven economic development.

At the same time, we can say that the implementation of smart cities raises significant challenges, such as cybersecurity, governance, financial sustainability and adaptation to climate change. However, they open new horizons through their opportunities such as predictive cities based on AI, urban simulations using digital twins, collaborative economies integrated on blockchain, public–private partnerships and investments in digital education. Success is an accumulation of variables that depends on the balance between innovation and legal regulations.

This work is not merely a review but a detailed analysis of current studies in the field and, at the same time, a conceptual system proposal, based on the research carried out and on the selection of the most efficient methods of approach. Thus, it outlines clear lines of further development through methodological, architectural and technological approaches conducive to the development of an integrated urban informational hub. It is important to note that the research is not oriented towards actual implementation and does not include relevant tests and analyses, because the orientation of this work is to present, in a condensed form, the multitude of studies in the field of smart cities, and through those literature analyses carried out an architectural system is also proposed that can be realised in the future.

The proposed prototype demonstrates that an integrated platform can facilitate the modernisation of urban services and improve the quality of life of citizens. The chosen design of the platform’s exposure is conceived in a modular way, being based on microservices and Kubernetes, offering technical feasibility, but at this stage it remains only a prototype presentation, not having been fully implemented and tested in real municipal infrastructures.

At the same time, the major difference between this work and what has been researched in other studies is that this research is oriented towards the idea of developing a unified system that allows the management of multiple activities from a single starting point. Compared with other systems that address only a certain area of interest, this system concept aims to unify the multitude of areas of an everyday city, making urban accessibility much easier to access. It is understandable that such a system requires very high costs, but this project is not isolated, since there are examples that are used even today in China, available in the WeChat/Alipay applications, which start as main applications that host multiple services/microservices for different areas of interest of the citizen. However, those applications are mainly intended for services that can be purchased, while the system proposal within this work is oriented more towards the administrative side of the city and facilitates the processes through which a citizen can access urban facilities.

In essence, the system aims to approach a modularisation technique so that microservices can be gradually added to the hub-type system.

The limitations of this proposed system are related to the fact that it is only at the stage of a conceptual prototype, industrial scalability is unvalidated, integration with municipal systems is not achieved, interoperability problems may be encountered, there is a high cost of real implementation and there are limitations related to data security, since it will have to be constantly strengthened to protect citizens from cyberattacks.

From a perspective of future developments, this conceptual prototype can be implemented at a level that allows a relevant evaluation and exposes simulated or even real scenarios in the case that access can be obtained to testing in a real environment.

Smart cities of the future are not only hyperconnected spaces but people-centred digital communities where technology, governance and sustainability harmonise.