The advancement of smart city services requires innovative approaches to manage urban transportation and enhance tourist mobility through sensor-based IoT data models for efficient processing. This work proposes an integrated tourist recommendation system within a multimodal travel orchestration platform, named Multipres. Designed to support sustainable tourism, Multipres optimizes urban infrastructure, including buses, bicycles, and pedestrian paths, while reducing resource strain. The platform leverages AI-based recommendation engines, integrates IoT mobility sensors, bike-sharing stations and vehicles’ parking platforms, and external applications, such as Waze. Results demonstrate that Multipres provides recommendations according to the users’ needs with updated information through the GenAI integration, in a multimodal approach with real-time data from the city sensing and external platforms, both for mobility and traffic events, and taking into account environmentally indicators.

Federated Learning (FL) is a promising approach for parameter normalisation in Machine Learning (ML) models, especially when data privacy and computing distribution are crucial. However, there are significant constraints in FL solutions, particularly concerning the handling of the mobility of participating nodes in the parameter aggregation processes, with a substantial impact on Vehicle to Everything (V2X) scenarios within the scope of smart cities. To address this challenge, we propose Mobile Federated Learning System (MobFedLS), a lightweight microservices-based framework capable of operating on various types of devices (mobile and non-mobile). MobFedLS features an interface to integrate ML models to cooperate in the FL process without intrusion between the parties. MobFedLS manages the entire federation process, from instantiating services on mobile nodes to the final parameter updates in the involved ML models and the release of resources used in all participating nodes. Additionally, MobFedLS handles node mobility and ensures the proper execution of federated processes, even with nodes entering and leaving at any stage of the aggregation process. To demonstrate the capabilities of MobFedLS, we use data collected through the city-scale infrastructure of Aveiro Tech City Living Lab (ATCLL), specifically the position of vehicles during their movement through the city. In the tests, we evaluate all phases of the aggregation process for mobile nodes. The results show that, even with intermittent connectivity to the city-infrastructure ATCLL, the MobFedLS system manages the node mobility and effectively handles node availability during the aggregation of ML model parameters.

Microservices offer advantages such as better fault isolation, smaller and faster deployments, scalability, and speeding up the development of new applications through the composition of services. However, its large-scale use and specific requirements increase the challenges of monitoring and management. To meet these challenges, we propose a monitoring and management system for microservices, containers and container clusters that autonomously predicts load variations and resource scarcity, which is capable of making new resources available in order to ensure the continuity of the process without interruptions. Our solution’s architecture allows for customizable service-specific metrics that are used by the load predictor to anticipate resource consumption peaks and proactively allocate them. In addition, our management system, by identifying/predicting low demand, frees up resources making the system more efficient. We evaluated our service management solution in the AWS environment, environment characterized by high mobility, dynamic topologies caused by disconnection, dropped packets and delay issues. Our results show that our solution improves the efficiency of escalation policies, and reduces response time by improving the QoS/QoE of the system.

The Federated Learning (FL) paradigm aims to provide performance advantages over centralized models, such as lower latency and communication overhead when doing most of the processing on the edge devices, better privacy as data does not travel over the network, easier handling in heterogeneous data sources and better scalability. However, the development of FL-based solutions is done through tools aimed for specialists as it always requires some programming. To cover this gap, we present an architecture for a lightweight container-based solution that offers a range of machine learning (ML) algorithms to build prediction engines for edge devices, which also includes the main options in algorithms/models for aggregation and refinement of models in the central server. The proposed framework allows the rapid build of containerized testbeds for the evaluation of ML and aggregation algorithms in the initial evaluation phase, and also later in the installation in real production infrastructures. We demonstrate the efficiency of our approach in estimating vehicle mobility in and out of the city of Aveiro, using real data collected by the communications and sensing infrastructure.

This article presents the deployment and experimentation architecture of the Aveiro Tech City Living Lab (ATCLL) in Aveiro, Portugal. This platform comprises a large number of Internet of Things (IoT) devices with communication, sensing, and computing capabilities. The communication infrastructure, built on fiber and millimeter-wave (mmWave) links, integrates a communication network with radio terminals [WiFi, ITS-G5, cellular vehicular-to-everything, 5G and LoRa(WAN)], multiprotocol, spread throughout 44 connected points of access in the city. Additionally, public transportation has also been equipped with communication and sensing units. All these points combine and interconnect a set of sensors, such as mobility (radars, light detection and rangings (LiDARs), and video cameras) and environmental sensors. Combining edge computing and cloud management to deploy the services and manage the platform, and a data platform to gather and process the data, the living lab supports a wide range of services and applications: IoT, intelligent transport systems (ITSs) and assisted driving, environmental monitoring, emergency and safety, and among others. This article describes the architecture, implementation, and deployment to make the overall platform to work and integrate researchers and citizens. Moreover, it showcases some examples of the performance metrics achieved in the city infrastructure, the data that can be collected, visualized, and used to build services and applications to the cities, and, finally, different use cases in the mobility and safety scenarios.

In recent years, the Internet of Things (IoT) has become a standard facet of modern communications, and information-centric networks have been pointed as an alternative to bypass the restrictions imposed by the traditional Internet protocol networks regarding the mobility of its network elements. However, the improvements imposed by this new paradigm fall short in large scale mobile wireless distributed environments inherent to IoT, due to high node mobility, dynamic topologies and intermittent connectivity. To tackle these issues, we present a named data network (NDN)-based publish–subscribe mechanism with support for both Consumer and Producer mobility. This approach handles the Producer mobility by combining the Data packets with infrastructure specific information, fixing the broken paths between the Producer and the Consumer; and the Consumer mobility by monitoring and anticipating mobile node trajectories while compelling the infrastructure to adjust to new paths. Simulation results, assuming a smart city use case and using real traces of vehicular mobility, have shown that the proposed solution far surpasses the native NDN workflow and traditional publish–subscribe solutions. With respect to the Producer mobility, the proposed solution delivers 79% of Data packets against 14% with the Native implementation, when using 25 mobile Producers; regarding the Consumer mobility, results have shown that our solution achieves almost the same Consumer satisfaction ratio as previous implementations but reducing substantially the network overhead related with the transmission of Interest packets.

In 5G networks, time-series data will be omnipresent for the monitoring and management of network performance metrics. With the increase in the number of Internet of Things (IoT) devices, it is expected that the number of real-time time-series data streams will increase at a fast pace, making forecasting essential for the proactive successful management of the network. In this article, we discuss to use both linear and non-linear forecasting methods, including machine learning, deep learning, and neural networks to improve 5G networks’ management. For this purpose, we design and implement a real-time distributed forecasting framework, used to make simultaneous predictions of different network performance metrics, and with different learning algorithms. By using our framework, we compare the use of forecasting methods in two network scenarios, in a real vehicular network and in a 4G network, representing two different slices in a 5G network. We also integrate our framework in a 5G architecture. Using the best forecasting models assessed previously, we propose a dynamic threshold algorithm for multi-slice management, to ensure that the resources of each slice are updated according to the slices’ needs, while avoiding congestion and saving resources for other slices. The experimental results show that it is possible to forecast the slices’ needs and congestion probability, selecting the best forecasting approach or an ensemble of the best ones, and act accordingly in the network to optimize its management.

Emergency services play an important role in the intelligent transportation systems based on mobile communication networks in smart cities. However, the characteristics of Vehicular Ad-hoc Networks (VANETs), such as high mobility, intermittent connectivity, scalability and constant changes in the network topology make this type of message dissemination a challenge. To improve emergency message dissemination between vehicles, we propose strategies that take advantage of the location, direction, speed, number of vehicles’ neighbors and characteristics of the region of the city, so that the message reaches all vehicles in the shortest time with the lowest network overhead. To show the effectiveness of our strategies, we deployed an emulation platform with real connectivity and mobility data. Our results show a delivery rate between 92% and 100% at various densities and vehicle speeds, for various zone sizes of relevance and range of the communication technology.

For efficient network management, it is important to monitor and analyse the data, particularly big data applications based on time series, in terms of trends and correlations, to predict network problems and be able to react preventively. Machine learning techniques can help but, given the amount and complexity of the algorithms and metrics available, the use of these techniques is laborious and requires specialized knowledge. This paper proposes a framework for distributed real-time time series forecasting with the goal to make predictions for various dynamic systems simultaneously and provide straightforward horizontal scaling, increased modularity, high robustness and a simple interface for users. Moreover, the proposed framework also enables the creation of ensemble algorithms, combining the results of multiple predictors, without changes on each individual predictor component. To demonstrate the functionalities of our framework, we show how simultaneous predictions can be made about the number of Internet sessions, using a real data stream from users of the buses in the Porto city.

Content distribution strategies for delayed-tolerance networks (DTNs), such as Local Rarest Bundle First (LRBF), reach high delivery rate in the vehicular ad-hoc networks (VANETs), in spite of their continuous node mobility and their wide geographical dispersion. However, to reach a high delivery rate, this strategy requires each node to periodically send control packets with information of which content bundles it contains and that are available to its neighbours, increasing the network overhead. To overcome this problem, we are proposing two content distribution approaches to minimize the size of the control packets. Our results show that our algorithms can reduce more than 90% the size of control packets, while decreasing the network overhead without changing the delivery rate and end-to-end delay.

A common concern among Smart Cities is the focus on sensing procedures to provide city-wide information to city managers and citizens. To meet the growing demands of Smart Cities, the network must provide the ability to handle a large number of mobile sensors/devices, with high heterogeneity and unpredictable mobility, by collecting and delivering the sensed in- formation for future treatment. With these challenges in mind, we are proposing self-organizing opportunistic approaches that allow vehicles to disseminate information to each other even while on the move, over a multi-technology opportunistic communication platform for data gathering and data exchange with both static and moving elements. To demonstrate their characteristics, we perform real experiments to evaluate four distinguished routing strategies, Epidemic, Direct Contact, Controlled Replication with Neighborhood Classification - Contacts based and Controlled Replication with Neighborhood Classification - Mobility based over an opportunistic vehicular network.

The Local Rarest Bundle First (LRBF) and Local Rarest Generation First (LRGF), content distribution strategies for delayed-tolerance networks (DTNs), reach high delivery rate in the vehicular ad-hoc networks (VANETs), where intermittent connectivity due to permanent network fragmentation and node mobility are main challenges. However, in order to achieve a high delivery rate, both strategies require each node to periodically transmit control packets with information about the availability of its resources to its neighbors. This large amount of control packets causes overhead on the network. To overcome this effect, we are proposing a content distribution strategy based on a Bloom Filter. The results showed that our strategy significantly reduced the size of control packets decreasing the network overhead. Besides that, the number of nodes that obtains the least disseminated packets is much smaller.

MapReduce is a programming model for processing and generating large data sets, and Hadoop, a MapReduce implementation, is a good tool to handle Big Data. Cloud computing with its ubiquitous characteristic, on demand and dynamic resource provisioning at low cost has potential to be the environment to treat big data. However, using Hadoop on the cloud spends time and requires technical knowledge from users. The hybrid cloud leverages these requirements, because it's necessary to evaluate the resources in private cloud and, if necessary, obtain and prepare on-demand resources in the public cloud. Moreover, the simultaneous management of private and public domains requires an appropriate model that combines performance with minimal cost. In this paper we propose an architecture to make the orchestration of Hadoop applications in hybrid clouds. The core of the model consists of a web portal for submissions, an orchestration engine and an execution services factory. Through these three components it's possible to automate the preparation of a cross-domain cluster, performing the provisioning of files involved, managing the execution of the application, and making the results available to the user.

Cloud computing has recently emerged as a convergence of concepts such as cluster computing, grid computing, utility computing, and virtualization. In hybrid clouds, the user has its private cloud available for use, but she can also request new resources to public clouds in a pay-per-use basis when there is an increase in demand. In this scenario it is important to decide when and how to request these new resources to satisfy deadlines and/or to get a reasonable execution time, while minimizing the monetary costs involved. In this paper we propose a strategy to schedule service workflows in a hybrid cloud. The strategy aims at determining which services should use paid resources and what kind of resource should be requested to the cloud in order to minimize costs and meet deadlines. Experiments suggest that the strategy can decrease the execution costs while maintaining reasonable execution times.