The integration of Internet of Things (IoT) networks with satellite communication systems is increasingly gaining attention for enabling connectivity in remote and infrastructure-limited environments. However, the experimental validation of heterogeneous IoT-satellite communication architectures remains challenging due to the complexity and cost associated with testing infrastructures. This work presents the design and implementation of a modular laboratory testbed that enables controlled evaluation of IoT-CubeSat communication chains using commercial off-the-shelf components. The proposed architecture integrates LoRa-based terrestrial IoT nodes with a nanosatellite communication segment through a gateway platform equipped with a UHF satellite modem. An experimental methodology based on closed-loop communication with a nanosatellite engineering model is implemented to validate end-to-end data transmission across the heterogeneous links. The results demonstrate successful packet generation, forwarding, satellite transmission, and reception, confirming the functional interoperability of the proposed system. The developed platform provides a flexible and cost-effective solution for rapid prototyping and validation of IoT-satellite communication concepts in laboratory environments, supporting future research on hybrid terrestrial-space communication architectures.

Nanosatellite operations increasingly rely on telecommand scripts to automate payload activities, communication sessions, and subsystem mode transitions. Although scripting improves operational efficiency, it complicates power planning because a single script (or short command sequence) may activate multiple regulated Electrical Power System (EPS) outputs and drive concurrent current draw. This paper proposes an NLP-based multi-label prediction framework that infers which EPS outputs become active directly from telecommand script text. The dataset is built from two 3U nanosatellites operated by the University Center for Research in Space Technology at Mohammed V University (Rabat) and combines (i) ground-segment task logs containing both scheduled scripts uploaded for execution outside contact windows and in-pass scripts executed during ground-station visibility, and (ii) EPS telemetry streams reporting timestamped per-output currents. Training instances are created by temporally aligning script execution to telemetry intervals; labels are generated using an idle-aware activation rule that classifies an output as active when its post-interval current exceeds an empirically characterized idle level plus a noise margin. We compare a TF-IDF + MLP baseline against a transformer encoder adapted to command-like domain-specific language. On the held-out test set, the TF-IDF baseline achieves 0.65 micro-F1, and the transformer model achieves 0.91 micro-F1, showing that command semantics captured by NLP can provide operationally meaningful power awareness for planning and verification.

The rapid advancement of satellite technologies driven by miniaturization, automation, and cost-effective manufacturing has a significant impact on the role of satellite systems in critical infrastructure and global communications. However, this expansion accompanies a growing array of security threats, particularly at the application layer of satellite communication protocols such as the CCSDS (Consultative Committee for Space Data Systems) File Delivery Protocol (CFDP). This paper proposes a novel security architecture that integrates a hybrid cryptographic framework combining Advanced Encryption Standard and Elliptic Curve Cryptography within CFDP. This integration ensures end-to-end confidentiality, integrity, and authentication of file transfers between satellites and ground stations. In doing so, the paper addresses emerging vulnerabilities in current security architectures. The proposed design further enhances system resilience through GPS-based time synchronization and a dual-band UHF/S-band communication strategy. A comprehensive threat modeling analysis using the STRIDE framework systematically identifies and mitigates risks such as spoofing, tampering, repudiation, information disclosure, and elevation of privileges. Comparative evaluations demonstrate the robustness of the proposed architecture in countering cyber intrusions and safeguarding application-layer data integrity, while also highlighting persistent challenges related to physical-layer attacks such as jamming. The results establish a robust, multi-layered security framework adapted to the specific operational constraints of modern satellite systems, offering valuable insights for future research aimed at strengthening the cybersecurity posture of space-based networks.

Client selection remains a critical challenge in Federated Learning (FL). Resource-aware strategies aim to reduce training delays and mitigate stragglers by selecting an appropriate subset of clients in each round. However, these methods prioritize computationally strong clients and exclude resource-constrained clients. In healthcare settings, this approach is impractical because it removes entire domains from training, which harms generalisation. To address these challenges, we propose CDCSF, a domain-aware client selection framework that re-partitions clients into domain-homogeneous groups in each iteration. CDCSF is a dynamic clustering framework based on the (EM) algorithm that clusters clients based on local feature prototypes to enhance domain diversity. The framework incorporates a reliability score derived from an exponential moving average of training time to favor efficient clients. Simultaneously, a fairness score is introduced to ensure that underrepresented clients can still contribute to the training. This approach preserves sufficient representation across all domains to improve model generalization and accelerate convergence. We conduct extensive experiments on a healthcare benchmark dataset to validate the effectiveness of CDCSF. The proposed method improves accuracy by 2% over FedAvg under domain shift and outperforms PoC by 8%. With the proposed adaptive client selection strategy, we further demonstrate that CDCSF converges significantly faster than baseline methods under heterogeneous resource and data conditions.

CubeSats have become an important platform for low-cost space missions by integrating multiple tightly coupled subsystems whose interactions strongly affect mission performance and reliability. These nanosatellites continuously generate massive volumes of telemetry data that are used to monitor subsystems health. Most existing studies analyze telemetry data at the level of individual subsystems, which limits the understanding of system-level interactions and dependencies. This paper proposes a data-driven framework to investigate inter-subsystem relationships in CubeSat platforms using telemetry data. The approach first applies a threshold-based Out-Of-Limits anomaly detection method to identify abnormal operational periods. Temporal windowing is then performed around detected anomalies, followed by multi-subsystem data fusion. Cross-subsystem dependencies are quantified using the Pearson correlation coefficient, enabling the analysis of statistical relationships between telemetry parameters from different subsystems. The framework is evaluated using telemetry data from five CubeSat subsystems: Electrical Power System, Attitude Determination and Control System, Communications, Flight Computer, and Payload Controller. The results show that meaningful correlation patterns emerge during abnormal operating conditions, particularly between the electrical Power System and the attitude Determination and Control System. The temporal behavior of these correlations highlights dynamic interaction mechanisms within the CubeSat architecture. The proposed framework offers a lightweight and interpretable approach for system-level telemetry analysis and provides a structured basis for building well-prepared datasets for future anomaly detection models without relying on detailed physical models or labeled fault data.

This article presents a real-time testing platform designed to evaluate the timing behavior and communication reliability of CubeSat subsystems. At its core, the platform integrates a custom logger, along with mock software modules that simulate key CubeSat subsystems. These modules are implemented using the Zephyr real-time operating system (OS) and communicate using the CubeSat space protocol (CSP) over a controller area network bus. The system supports controlled and repeatable experimentation, enabling precise assessment of message latency, deadline adherence, and intertask communication dynamics. In addition, the platform offers extensible interfaces to support the integration and testing of flight software components under realistic operational constraints. The proposed platform facilitates early stage design validation and performance characterization within the CubeSat development lifecycle, offering deeper insights into intersubsystem communication behavior. This enables development teams to deliver more reliable software and produce more accurate data link budgets, with particular emphasis on timing determinism and subsystem interaction fidelity.

The use of IoT devices especially sensors and actuators continue to emerge across many domains: Smart city, Smart vehicle, Smart healthcare, Smart Factory. . . and influences the way we live and act nowadays. Nevertheless, IoT networks are subject to many security concerns related to privacy, data integrity, traceability, reliability and more. Device authentication is the key stone to prevent security attacks, guarantee the efficiency of the IoT network and ensure a reliable information exchange. Many works have been done around device-to-device and user-to-device authentication protocols and different factors have been considered: RFI tags, MAC address, PUF, Unique ID, . . . all aiming to provide an efficient and secure authentication scheme that suits the heterogeneous resource constrained IoT devices. In this paper we discuss several ways of device-to-device authentication for IoT applications presented by researchers to ensure enhanced authentication protocols and key exchange mechanisms. We also highlight common challenges for these mechanisms in relation with timestamp accuracy and time synchronization which seems to be a green field for scientific researches related to IoT.

The CubeSat Space Protocol (CSP) has proven to be an effective and efficient communication protocol specifically designed for distributed embedded systems within small networks such as those found in nanosatellites. The lack of standardisation in this protocol has raised concerns among space engineers about interoperability, error rates and reliability. Despite these concerns, CSP has been widely adopted and integrated into several major projects. This paper highlights the critical need for standardisation of CSP, with a focus on robust modelling and verification methodologies for data transmission. The study provides an in-depth analysis of CSP, emphasising its client-server architecture through finite-state machine modelling. In particular, advanced verification techniques using the Promela language are employed to provide a high level of reliability for the protocol by rigorously examining packet transmission. The results confirm the effectiveness of CSP and highlight the need for standardisation, allowing CubeSat developers to implement the protocol with greater confidence and efficiency.

Satellite mission planning and optimization require precise scheduling to maximize the efficiency of communication and payload operations. This paper presents a machine learning (ML)-based framework for predicting data transmission durations for upload and download tasks during satellite passes. The framework leverages historical data from a network of two 3U satellite systems, UM5-EOSAT and UM5-RIBAT, and integrates real-world constraints including S-band antenna limitations, antenna coverage, elevation mask constraints, pass geometry, and the size of files to be downloaded. Feature engineering and model evaluation, including XGBoost, LightGBM, and Random Forest, reveal XGBoost’s superior predictive performance R2 = 96%, MAE= 10.25 seconds. This framework enhances scheduling efficiency and operational performance, contributing to more effective mission planning.

In the realm of space applications, software reliability plays a vital role in defining the success or failure of a mission. For Nanosatellite missions, software is typically developed on demand, tailored to meet the specific needs and constraints of the system. These constraints often necessitate a minimalistic approach to software design to ensure safety and secure operation. However, the software must also be sufficiently complex to achieve all mission objectives. While the use of general purpose operating systems like Linux has been increasing in the past years, commercial off-the-shelf real-time operating systems (RTOS) are still preferred due to their deterministic nature and the ease of development and debugging. This paper explores the integration of a FreeRTOS layer into an MSP430-based Electrical Power System for a nanosatellite. The integration involves adding telemetry collection and housekeeping tasks to enhance subsystem integration and responsiveness with other subsystems without putting pressure in the system resources.

As artificial intelligence (AI) algorithms continue to advance, researchers have leveraged deep neural networks to address a range of challenges in the medical field. These models require a large-scale dataset and high-quality annotated data for model generalization, which is a major challenge in imaging data due to their limited availability in healthcare institutions. Additionally, it is primarily challenging to work with private patient data and share it with an external entity due to the privacy concerns. These challenges of the traditional centralized learning have led to a more efficient decentralized approach. This approach involves training with a diverse range of data from various domains, which are required to enhance model performance. Hence, many researchers have adopted Federated learning as an emerging paradigm to collaboratively train a machine learning model among multiple healthcare institutions without sharing their local private data. However, medical datasets are sourced from different medical institutions; hence they are often acquired by different protocols, scanner types, data modalities, and from different patient populations. Thus, it is inherently heterogeneous which degrades the global model performance in the federated setting. In this paper, we explore the key motivation for using federated learning in the healthcare field and discuss the challenges posed by the diversity and data heterogeneity of medical data from various institutions. Additionally, we present recent works that help mitigate the non-iid data issue in federated learning. Furthermore, we empirically evaluate the federated learning algorithms alongside centralized learning and one site learning using a benchmark medical dataset. Our analysis demonstrates that the adoption of advanced methods in FL enables us to effectively mitigate the data heterogeneity issue while leveraging data privacy and large-scale datasets within the medical domain.

The rapid proliferation of mobile devices and Internet of Things (IoT) gadgets has led to a critical shortage of spectral resources. Cognitive Radio (CR) emerges as a propitious technology to tackle this issue by enabling the opportunistic use of underexploited frequency bands. Automatic Modulation Classification (AMC), which serves as a technique to blindly identify modulation types of received signals, plays a pivotal role in carrying out several CR functions, including inference detection and link adaptation. Recent research has turned to Deep Learning (DL) networks to overcome the shortcomings of traditional AMC techniques. However, most existing DL approaches are impractical for resource-limited systems. To address this challenge, we propose a novel lightweight hybrid neural network for AMC that fuses Convolutional Neural Networks (CNNs) and Gated Recurrent Units (GRUs) layers, along with a customized Squeeze and Excitation (SE) block. The integration of CNNs and GRUs allows for the learning of both spatial and temporal dependencies in modulated signals, while the SE block recalibrates features by modeling interdependencies between CNN network channels. Our experimental results, using the RadioML 2016.10A dataset, clearly demonstrate the superior performance of our approach in effectively managing the tradeoff between accuracy and complexity compared to baseline methods. Specifically, our approach achieves the highest accuracy of 91.73%, surpassing all reference models while reducing the memory footprint by at least 45%. In future work, further investigation is warranted to differentiate modulations sharing temporal or frequency domain characteristics and enhance classification accuracy in high-noise environments.

Automatic Modulation Classification (AMC) is a fundamental task to blindly identify modulation schemes within Radio Frequency (RF) signals. This process holds paramount importance within the realm of Cognitive Radio (CR) applications, enabling critical tasks such as interference detection and link adaptation. Deep learning (DL) has shown remarkable effectiveness in addressing the AMC challenge; however, most DL-based AMC schemes have high processing and storage requirements, rendering them unsuitable for resource-constrained devices. To tackle this issue, this paper introduces a lightweight neural network (NN) constructed by fusing Gated Recurrent Units (GRUs) and multiple convolutional blocks. These convolutional blocks are meticulously designed using asymmetric kernels to reduce computational complexity, and Squeeze and Excitation (SE) blocks to enhance channel interdependencies with minimal computational cost. Additionally, the convolutional blocks incorporate skip connections to improve classification accuracy and mitigate the vanishing gradient problem. Experimental results on the RadioML 2016.10A dataset demonstrate that our model outperforms benchmark DL-based AMC models in terms of accuracy while utilizing the fewest trainable parameters and exhibiting moderate inference time.

As part of the significant evolution that the space industry is experiencing, a fast increase in the number of CubeSats projects for scientific, commercial and military purposes has been noted in recent years. This acceleration, coupled with the widespread use of Commercial Off-The-Shelf (COTS) components, raises questions about the ability of these systems to withstand potential cyberattacks, which are becoming more prevalent. Thus, the cyber resilience of a CubeSat depends on its ability to effectively detect attacks despite the constraints of autonomy and the limitation of resources that characterize the space missions. To address this need, our paper proposes an Intrusion Detection System (IDS) for CubeSat systems. This distributed solution uses an Artificial Neural Network (ANN) module for classifying CSP packets over CAN on board the space segment based respectively on timestamp and Data field, while the classifier training processes are executed at the ground segment level. The results obtained following the experimentation of this IDS against three types of common attacks are very encouraging thanks to detection rates obtained between 87.66% and 99.59% (F1-score).

The increasing dependence on satellite technology for critical applications, such as telecommunications, Earth observation, and navigation, underscores the need for robust security measures to safeguard these assets from potential cyber threats. Moreover, as many satellite systems rely on the Controller Area Network (CAN) protocol for efficient data exchange among onboard subsystems, they become prime targets for cyberattacks. While contributions present various options for detecting attacks in the CAN bus, no one proposes an architecture suitable for satellite systems. To address this concern, this paper presents a novel approach to develop an adaptive distributed Intrusion Detection System (IDS) for satellites, which integrates machine and deep learning techniques for the classification of CAN frames. This system is specifically designed to overcome the inherent power and computational challenges of satellite operations by executing time-based anomaly detection on board, and content-based detection at the ground segment. To evaluate the effectiveness of the proposed solution, experiments are conducted using representative Datasets. The obtained results demonstrate that the distributed IDS presented in this research offers a promising solution to improve the security of satellite systems by achieving high detection rates ranging from 91.12% to 99.86% (F1-score).

The current paper presents a clustering method for energy efficient management of heterogeneous nodes of a flying ad hoc network (FANET). The technological advances of the last decade gave rise to emerging technologies. Unmanned aerial vehicles (UAVs) are small aircraft that proved their usefulness for different tasks nowadays. They can collaborate to achieve missions especially in areas where traditional networks cannot work or cannot accede. A FANET is composed by a number of these aircraft. For alike networks, the resources are limited. Indeed, an efficient energy management is required to extend the life of the network. This work is a clustering method for heterogeneous nodes of a FANET, each node is equipped with one sensor, and four different sensors are used. Clustering is grouping nodes with the aim of efficiency improvement. The clustering is done before the beginning of the rescue mission and depends on the types of sensors the nodes are equipped with and. The master election depends on the available energy of each one of the nodes. The simulation is done with a discrete event simulator (DES) and the results are compared to the algorithm of glowworm swarm optimization (GSO) to demonstrate the effectiveness of the suggested technique.

The impressive growth in the number of CubeSats projects carried out over the past decade gives rise to a new segment in the space industry. Whether academic, military or commercial, CubeSats missions provide a wide range of low-cost space applications. This is due on the one hand to the evolution in the launch offer and on the other hand to the use of Components-Of-The-Shelf (COTS). Despite their advantages, these components are associated with many cybersecurity risks, making space segments more accessible targets for cyberattacks and intrusions. This paper provides a brief overview of intrusion detection approaches corresponding to cyber threats that may target CubeSats. We classify Intrusion Detection System (IDS) components according to a number of factors, such as data sources, architectures, detection mechanisms and modes, and intrusion forms. Then, we present challenges and potential areas for future research, with the goal of designing a CubeSat Intrusion Detection System that efficiently balances security performance and space missions’ constraints.

The accelerated evolution in space industry has led to significant changes in the way space projects are conceptualized and executed. With greater reliance on the use of Commercial Off-The-Shelf and terrestrial technologies, space missions have become a major target for cyberattacks. To face these risks, it is crucial to enhance the cyber resilience of space systems through secure architectures that are tailored to the specific needs of each mission and based on a wide range of cyber threats. This article addresses this challenge by providing a holistic approach to building a Defensible Cybersecurity Architecture for New Space Missions. The five-step approach is based on understanding the mission context and (by extension) carrying out a risk analysis, to develop a mission-specific threat model using a cyber threat knowledge base. Within the framework of the generated model, the security architecture is implemented through a prioritized and layered application of security countermeasures according to "Zero-Trust" principles. Finally, monitoring and maintaining are carried out through the collection and analysis of security data, to continuously improve the relevance of the design. In addition, the application of the proposed approach on a FlatSat emulating an Earth-Observation CubeSat demonstrates its effectiveness in securing satellite projects.

Addressing the spectrum scarcity problem under the explosive growth of wireless network applications and data traffic, requires a decidedly global research focus on Cognitive Radio Networks (CRNs), which have been widely regarded as the front-runner to cope with spectrum inefficiency. In this paper, the emphasis is given to Automatic Modulation Classification (AMC) as a means to blindly identify the underlying modulation format of intercepted signals in Cognitive Radio Networks. Modulation recognition is a necessary task to accomplish a set of Cognitive Radio functions, such as dynamic spectrum access and interference detection. Recently, several Deep Learning (DL) schemes have been explored to address the existing limitations of conventional approaches in dealing with the AMC problem. DL techniques have delivered outstanding performance in terms of reliability, scalability, and applicability compared to traditional methods. However, when it comes to resource-constrained end-devices, most of these techniques are not supported due to their high processing and storage requirements. In this work, we propose a hybrid DL model with attention mechanism based on layers of bidirectional gated recurrent unit (BiGRU) and convolutional neural network (CNN) architectures. Our model has a memory footprint of 126.6 kByte, slightly smaller than the memory footprint of the baseline model with 100 GRU cells (128 kByte), and achieves good performance on different scales of training data.

In recent years, IoT devices have been widely used in a variety of sectors such as industry, smart farming, and smart homes. Its application requires performing high computational analysis in real-time. The research era of Artificial Intelligence has witnessed an intense development conducted by millions of research and applications that extend from systems recommendation to video/audio surveillance. AI algorithms have been deployed to IoT data to bring intelligent decisions for IoT applications. These numerous data increase the time of the data transition to the cloud, which becomes the bottleneck of the cloud-based architecture. The edge computing technology brings the AI algorithms to the Edge of the network to improve latency, bandwidth, and data privacy, and guarantee the high accuracy of the AI algorithms. Recently Federated learning (FL) is a machine learning technique that distributes the training among edge devices near to the data source in light of increasing privacy and leveraging from the massive data distributed among numerous edge devices. Therefore, in this paper, we introduce recent research that demonstrates the effectiveness of this approach and present the architectures, models, and methods that implement FL with IoT devices.