Global warming has caused significant thermal changes worldwide, increasing electricity demand by 20% to 30% in recent years. This surge disrupts the balance between energy production and consumption, sometimes leading to power outages or blackouts in certain countries. To mitigate this issue, optimizing HVAC systems, which account for more than 40% of global energy consumption, is a key solution, relying on either traditional or AI-based approaches. Our study examines the impact of gender, particularly women’s responses, on thermal comfort prediction in industrial environments with a high female presence. Using the ASHRAE RP-884 and Scales Project datasets and applying the XGBoost model, we demonstrated that women’s responses achieve the highest prediction accuracy at 68.60%, compared to 62.01% for men and 57.97% for combined responses. Moreover, based on the analysis of responses from different genders in the dataset excerpt, the comfort range for men falls within that of women, suggesting that integrating only female responses into thermal comfort models could enhance HVAC control, productivity, and energy efficiency.

Transfer learning is considered an effective technique that enhances model performance by using knowledge acquired from a source dataset to tackle a similar task on a target dataset. This technique is particularly valuable in fields where labeled data is limited, such as thermal comfort prediction. In the Tunisian context, the lack of specific thermal comfort data for office spaces occupied by multiple people represents a major challenge to optimizing workplace environments. To address this, we introduce a transfer learning-based approach using ASHRAE RP-884 data from countries with similar climatic conditions. In fact, these data were purified based on climate conditions and selected for office-type buildings. Three transfer learning methods were evaluated using three models: a Deep GRU-BiGRU model, a Deep GRU model, and a BiGRU model. Our results present that the transfer learning approach based on the Deep GRU-BiGRU model achieves the highest accuracy, reaching 66.15% in thermal comfort prediction, outperforming the other methods.

Assessing student engagement and performance in online learning environments is essential for advancing adaptive tutoring systems. This study examines how question formulation influences student response behavior by analyzing response time, help feature usage, and accuracy. An initial online test with a formulation error led to longer response times and greater reliance on help features, indicating comprehension difficulties. A second test with a revised question format, involving 60 students, resulted in shorter response times, reduced help usage, and higher accuracy. These findings provide empirical evidence that unclear assessment items can increase cognitive load, hinder performance, and distort evaluation outcomes. Systematically validating question design can improve student engagement and learning effectiveness. Future research should explore AI-based tools to automatically detect and correct formulation issues in digital education environments.

International audience

In this work, a novel interval-based/data-driven safety verification technique is introduced for Intelli-gent/Autonomous Vehicles (I/AV). The interval arithmetic is adopted to enhance the reliability of the analytical models used for the autonomous navigation. Furthermore, a data-driven technique, which monitors the correlation relating variables of the modeled system, is adopted to ameliorate the uncertainty assessment. In such a manner, tight bounds of safety margins are obtained. To provide reliable safety verification, the proposed risk management approach has been integrated on an Adaptive Cruise Control (ACC) system. It permits to detect erroneous uncertainty estimation of an Extended Kalman Filter (EKF). Simulation results prove the overall risk management efficiency and its ability to handle uncertainties.

The current Internet is facing many challenges to accommodate the growing demands of cloud computing. Software‐defined networking (SDN) is an emerging paradigm, proposed to be used in this context in order to investigate these issues. SDN virtualises network functions, reducing the amount of hardware and allowing to control data centres through software and programs. With SDN, it is possible to build the future Internet and solve the current issues. SDN assures the programmability of the network components and the development of new routing and forwarding approaches without changing the hardware components. However, the transition from traditional Internet protocol (IP) networks to full SDN networks is associated with many challenges in a short term as it requires from network operators a costly investment on new networking equipments. So, fulfilling full SDN architectures must be incrementally introduced in conventional IP networks. This study presents a performance study of a hybrid IP/SDN network, under NS‐3, using Fat Tree, BCube and Three‐tier data centre network (DCN) architectures, compared to conventional IP. Simulation results show that using a hybrid IP/SDN network enhances the performance of DCN architectures by providing an efficient communication and minimising the energy consumption.

With the arrival of the Smart Grid, the electrical network is undergoing deep changes that has also included the data communication field. To study the behavior and the efficiency of this technology, a Power Line Communication (PLC) module based on the network simulator NS-3 was recently developed to compete with the usual Matlab evaluation. This allowed a more realistic simulation of PLC-based Smart Grid communication on all levels of the OSI model. Through this work, we present a comparative analysis between the PLC theoretical model, evaluated with Matlab, and the NS-3 dedicated module in a context of an outdoor communication. To compensate power losses over extended networks, we adopted the cooperative communication using a relay station at a residential city. The comparative study includes direct and relay-assisted transmission based on the Amplify and Forward (AF) and the Decode and Forward (DF) protocols, in the presence of several noises. The different implementations showed the performance of the cooperative communication in terms of the efficiency of the channel. Also, implementations proved the relative correspondence of the analytical results and simulation outputs for two different modelization approaches which are the ABCD matrix and voltage ratio.

This paper proposes a single switch open-circuit fault diagnosis approach for three-phase three-level NPC inverter based on conducted emissions. The proposed approach uses a simple electromagnetic interference (EMI) filter, placed between the DC source and the conversion system, to meter conducted interferences. A time-domain analysis of the voltages across the EMI filter resistors is, firstly, utilized to identify the two pair semiconductors among the two upper or the two lower switches. The Teager energy operator based algorithm is, then, applied. This latter proves the efficiency of this method for the identification of the faulty switch. The effectiveness of this pioneering approach is enhanced by simulation results.

Named Data Networking (NDN) is a new content centric architecture for the future Internet aiming to improve communication performance compared to coventional IP. One of the building bricks of NDN is data caching. Various caching strategies were proposed in the literature in order to improve NDN performance. But, it is still unclear which strategy performs better than the others in terms of efficient communication and energy consumption using DataCenter Networking (DCN) architectures.In this paper, we propose to evaluate a new cache replacement scheme based on Pseudo-LRU algorithm. We combine it with NDN inside a Fat Tree based DCN architecture. For that aim, we extended the ndnSIM module developped under NS-3 and ran simulations to evaluate the performance of our proposition. We also combined it with ProbCache placement scheme and compared it to conventional LRU scheme. We used evaluation metrics such as packet delay, hop count, cache hit rate and consumed energy. Simulation results show that our proposed Pseudo-LRU policy based outperforms LRU by reducing packet delay, hop count, consumed energy and increasing cache hit rate indepedently from cache size.

Experimenting large-scale communication systems is difficult and challenging especially when new networking solutions are to be tested. This difficulty was met during our work on the evaluation of the Content Centric Networking (CCN) benefits for the Smart Grid. On one hand, CCN is still an emerging networking paradigm aiming at resolving host-centric networking inconveniences. On the other hand, Smart Grid is the new generation of power production and distribution system requiring an adapted networking solution that fulfills its particular needs. Previous research works promoting the use of CCN to manage smart grid communications are based on simulations and analytical modeling. Hence, further experimentations of this solution in such context are lacking. The present work aims at filling this gap. Our goal is to emulate a content-centric smart grid communication system to be able to analyze, in practical way, CCN performance in this particular context. In this paper, a particular focus is granted to emulate content-centric routing and forwarding strategies. Comparing the performance of various combinations allows us to provide a reality check of content-centric deployment in a smart grid. Indeed, the following conclusions are drawn: better results were obtained with Link State Routing associated with Best Route forwarding strategy at the expense of Hyperbolic Routing/Adaptive SRTTcombination.

Providing the new power system with well-tailored communication architecture is a key factor to guarantee expected smart grid capabilities. Among many requirements, quality of service awareness is a relevant feature for the smart grid communication system. Indeed, various data flows are to be exchanged in order to ensure smart grid services, having each different resilience, bandwidth and latency thresholds. Proposing a communication architecture that is inherently quality of service aware for smart grids is, then, the main idea of the present research work. We intend to take advantage of emerging content-centric and software-defined networks whose principles are highly aligned with the smart grid communication system needs.

This paper presents a fault diagnosis system for a sophisticated automotive embedded system. The latter is a Smart Distance Keeping system (SDK) which is an advanced adaptative cruise control implemented in a truck. The proposed diagnosis process is applied in a decentralized manner. To provide local diagnosis to one of the SDK electronic control unit, the approach based on the vertices principal component analysis is selected. Without modeling the diagnosed system, this technique handles the imprecision into the sensors measurements.

With the evolution of the current use of the Internet and the volume of content exchanged through the web, huge datacenter networks (DCN) are deployed to take into account the diversification and mass of contents. In these datacenters, the architecture is an important component that has a great impact on its performance. In addition, the communication inside them is usually IP-based, which has been proven with several limitations in this context. A new communication paradigm allowing to surpass IP limitations is Content Centric Networking (CCN). This paper presents a comparative performance study, with NS-3, between CCN-based communication and IP-based communication using a specific DCN architecture called Fat Tree. The simulation results show that CCN performs better than IP using Fat Tree in terms of average packet delay and average throughput while varying the hosts' number. Moreover, we provide a performance analysis between various caching policies using the Fat Tree architecture and CCN-based communication.

Proper integration of renewable energy resources is one of the most revolutionary aspects of a smarter power grid. While green energy sources like wind and solar have been around for decades, what is lacking is the proper infrastructure to support their integration. In addition to electric infrastructure, a well-tailored communication solution is, in fact, needed to enable efficient green energy sources introduction into the smart grid. Content-centric networking is a novel communication paradigm with attractive features that has been explored for various use cases. In this research work, we revisit this paradigm for better renewable energy resources integration. First, a description of this application communication environment is presented in order to point out the inadequacy of current host-to-host communication pattern. Then, qualitative analysis will highlight the benefits of content-centric networking for renewable energy integration into the smart grid. A simulation study shows, in fact, how content-centric networking outperforms classical point-to-point communication solutions while managing required communications for renewable energy resources integration.