Resource
2026 EN
Francisca Soto Astorga · Jose Manuel Martinez · Noureddin M. Ibrahim
+1 more
Solar peer-to-peer (P2P) energy trading is a promising approach that allows prosumers and consumers to exchange energy directly, which reduces dependence on the centralized power grid. However, implementing such a P2P energy trading system requires careful consideration of hardware, software, and communication networks to ensure smooth and secure energy transactions between participants. Few research works have addressed the technical implementations of P2P energy trading system. Furthermore, there are no comparisons between different energy trading architectures and/or configurations. To address this gap, this work aims to design and implement a decentralized Solar P2P energy trading platform (hardware/software) that allows local energy trading between prosumers and consumers in a rural microgrid in Chile, aiming at maximizing the utilization of distributed energy resources. The proposed platform will benefit from recent advances in the Internet of Things (IoT), communication networks, cloud services, and blockchain technologies. A decentralized market architecture has been developed and implemented, in a laboratory environment, together with a comparative analysis with respect to communication network performance, considering different wired/wireless communication technologies. This work contributes towards the transition towards a decentralized energy market, facilitating the adoption of renewable energy sources, and empowering local communities.
Resource
2026 EN
Hakan Can · Ibrahim Yucedag
Depression poses a major global health challenge, while most computational detection approaches remain constrained by the 512-token input limit of standard Transformer models, restricting their ability to capture longitudinal user behavior. Moreover, aggressive text preprocessing may remove expressive cues that are informative for mental health assessment. This study presents a controlled user-level benchmarking framework for depression detection from social media text, examining the effects of preprocessing strategy, behavioral metadata integration, and context length.Atwo-stage experimental design is adopted: first, Aggressive Cleaning (AC) and Selective Cleaning (SC) are compared across four 512-token Transformer backbones (BERT, RoBERTa, DistilBERT, and MentalBERT), with and without metadata; second, Longformer and BigBird are evaluated at both 512 and 4096 tokens under the SC setting. The results showthat SC provides a more favorable context-preserving preprocessing strategy, while long-context modeling yields the strongest overall gains. The best-performing configuration, Longformer with 4096-token input and behavioral metadata fusion, achieves 90.36% accuracy, 87.15% F1, and 0.9667 AUROC. Additional t-SNE and SHAP analyses indicate that extended context improves the organization of user-level representations, while behavioral metadata provides a complementary rather than dominant contribution, with night posting ratio emerging as one of the most influential features. Overall, the findings show that context-preserving preprocessing and long-context sequence modeling are decisive factors for improving large-scale, text-centered social media-based depression detection.
Resource
2026 EN
Imad Eddine Gherbi · Ibrahim Alhamrouni · Younes Zahraoui
+2 more
The integration of Electric Vehicle Charging Stations (EVCSs) into distribution networks presents significant challenges, including increased power losses and voltage instability. This paper presents an optimization framework for the optimal placement of EVCSs in radial distribution systems with existing distributed generation. The objective is to minimize total active power loss over a 24-hour period while adhering to operational constraints, including bus voltage limits and line thermal capacities. A deterministic Non-Linear Programming (NLP) model is formulated and solved using the Interior Point Optimizer (IPOPT). The methodology is validated on the IEEE 33-bus and 69-bus networks, which incorporate fixed photovoltaic and biomass generation units. The proposed deterministic approach successfully identified optimal EVCS locations that significantly reduce system-wide power losses and improve voltage stability. To validate its reliability, the proposed method is benchmarked against three popular metaheuristic algorithms: Particle Swarm Optimization (PSO), Genetic Algorithm (GA), and Grey Wolf Optimizer (GWO). The results demonstrate that while the deterministic NLP framework entails a higher computational cost, it consistently outperforms all three metaheuristic methods in solution quality. This advantage is particularly pronounced in the larger 69-bus network, where the NLP model achieved a 47.15% reduction in power losses, whereas the metaheuristic algorithms failed to exceed 7.1% due to premature convergence. These findings indicate that, for standard radial distribution systems, the deterministic optimization framework offers a superior trade-off between solution optimality and reliability compared to stochastic metaheuristics, albeit at a higher computational cost. While the NLP model guarantees convergence to a high-quality solution, the study highlights the importance of selecting the appropriate optimization strategy based on the specific network size and available computational resources.
Resource
2026 EN
Caiwen Li · Iskandar Ishak · Hamidah Ibrahim
+3 more
The rapid expansion of e-commerce platforms has intensified demand for scalable, high-quality user segmentation systems capable of efficiently processing millions of behavioral records. This paper presents BIRCH-AE, a hierarchical ensemble clustering framework that integrates the Balanced Iterative Reducing and Clustering using Hierarchies (BIRCH) algorithm with autoencoder-based feature learning for large-scale e-commerce analytics. The autoencoder compresses high-dimensional behavioral data into compact latent representations, mitigating the curse of dimensionality and improving cluster separability. Multiple BIRCH configurations are combined through four ensemble strategies: Majority Voting, Weighted Voting, Advanced Affinity-based Spectral Clustering (AASC), and the proposed BIRCH-Optimized Hierarchical Consensus (BOHC). Dynamic selection based on multi-criteria evaluation automatically identifies the best-performing strategy per dataset setting, emphasizing that no single consensus method is universally optimal. Experiments on two large-scale datasets (Retail Rocket with 1.4M users and E-Commerce Behavior with 4.5M users) show improved clustering quality and practical scalability. BOHC achieves up to 23% silhouette improvement over single BIRCH for transaction-focused data with clearer hierarchical structure, while multi-domain data favor strong base models. Autoencoder feature learning improves clustering quality by 23–53% over raw features. The full 4.5M-user experiment was executed as a BOHC scalability run, completed in approximately 5 minutes, while framework-level comparative analyses were conducted through repeated stratified 30% subset trials. These findings support BIRCH-AE as a practical and adaptive segmentation framework for enterprise-scale e-commerce analytics.
Resource
2026 EN
Mohd Y. Ahmad · Bojan Petrovic · Igor Putnik
+5 more
This paper presents the design, fabrication, and performance evaluation of a flexible loop antenna integrated with LC (inductor-capacitor) sensors for wireless dielectric sensing in oral care applications. The antenna, implemented on a flexible substrate, enables conformal placement and reliable inductive coupling with passive LC sensors, facilitating non-invasive detection of dielectric property changes in biofluids. The flexible antenna was also tested at various bending angles. Difference in S11 magnitude for the bending angle of 0° and 40° was 2.5 dB with the magnitude being −17.5 dB for 0° and −15 dB for 40°. Two sensor types—big and small spiral laminated LC structures (BSL and SSL) were characterized under various test media, including deionized water, artificial saliva, and multiple concentrations of rosemary essential oil emulsions. Experimental results demonstrate stable resonance behavior under mechanical deformation and effective frequency shifts corresponding to liquid permittivity changes. Performance comparisons with conventional PCB and wire loop antennas confirm that the flexible design achieves comparable sensitivity with the frequency shift for all the antennas being around 100 MHz for BSL and 200 MHz for SSL in the liquids tested. The system exhibits strong linearity (R 2 > 0.98) in distinguishing environmental permittivity, highlighting its potential for future integration into wireless biosensing platforms for real-time, oral health diagnostics.
Resource
2026 EN
Hamza Ibrahim · Love Allen Chijioke Ahakonye · Jae-Min Lee
+1 more
Securing industrial Internet of Things (IIoT) systems presents critical challenges due to resource constraints, expanding attack surfaces, and the inadequacy of conventional security solutions against sophisticated cyber threats. While AI-driven detection and blockchain technologies offer promise, existing frameworks suffer from computational inefficiency, a lack of real-time prevention, or insufficient auditability. This article introduces a unified AI-PureChain intrusion prevention framework that tightly integrates deep learning-based threat detection with an immutable PureChain ledger using proof of authority and association (PoA 2 ) consensus. The proposed architecture achieves high-fidelity intrusion detection through hybrid CNN-BiLSTM models, attaining 99.76% accuracy on IoTForge Pro, 98.33% on WUSTL-IIoT-2021, and 98.11% on X-IIoTID datasets, while maintaining low inference latency (0.0016 s). The PureChain layer ensures tamper-proof audit trails with 24.56 transactions per second (TPS) throughput and 68-ms commit time, enabling verifiable prevention actions. Experimental results demonstrate complete attack mitigation (0% success rate) under high-traffic conditions while maintaining minimal resource consumption (14.49% CPU, 448-MB memory, 12.95-W power). This work represents a significant advancement in IIoT security by delivering a tightly coupled framework that simultaneously addresses detection accuracy, prevention reliability, and forensic accountability, thereby bridging critical gaps in current industrial security paradigms.
Resource
2026 EN
Hamza Ibrahim · Love Allen Chijioke Ahakonye · Jae-Min Lee
+1 more
The Industrial Internet of Things (IIoT) has transformed critical infrastructure but has also introduced severe security vulnerabilities, with breaches capable of causing catastrophic physical and operational damage. While blockchain technology offers a promising foundation for tamper-proof logging, existing platforms are often ill-suited for IIoT due to high latency, low throughput, and excessive energy consumption. Furthermore, most current research treats intrusion detection, secure logging, and system recovery as isolated components, lacking a unified framework for autonomous, verifiable resilience. To bridge this critical gap, this article introduces PureChain, a holistic, secure, and resilient ecosystem. PureChain integrates a custom lightweight blockchain with a deep learning-based intrusion detection system (IDS) and a novel verifiable recovery protocol, creating a closed-loop security model. The framework leverages a novel proof-of-authority and association (PoA 2 ) consensus mechanism, achieving high throughput (16.82 TPS), low latency (0.0594 s), and minimal energy consumption (12.43 W), demonstrating suitability for resource-constrained IIoT environments compared to general-purpose platforms like Ethereum and Hyperledger, which are optimized for different use cases. Upon intrusion detection by optimized models like XGBoost (99.87% accuracy), immutable blockchain logs actively trigger and cryptographically attest to infrastructure-enforced recovery actions such as device isolation via software-defined network (SDN) switches or state rollback through hardware management controllers. Extensive evaluation on benchmark IIoT datasets (IoT-CAD and IoTForge) demonstrates a detection-to-recovery success rate of up to 98.59% while maintaining 100% data integrity. PureChain establishes a new paradigm that unifies real-time threat intelligence, blockchain-based trust, and provable autonomous recovery for next-generation IIoT security.
Resource
2026 EN
Chun Wang · Hans Herdian · Wenbin Zheng
+16 more
This article describes a 300-GHz-band four-element amplifier-last phased-array transmitter (TX) that has been implemented in a 65-nm CMOS technology. Each chip integrates four TX elements with a power amplifier (PA)-last architecture. The proposed wideband PA employs an optimized transistor layout to reduce parasitic parameters, enhancing the transistor-gain corner frequency from 250 to 300 GHz for an $8~\mu $ m$\times 60$ nm device. A dual-peak $G_{\max }$ -core topology is adopted to achieve wideband inter-stage conjugate impedance matching. Operating from 237 to 267 GHz, the PA provides over 20 dB of gain while eliminating the need for power combining, thereby improving both TX output power and area efficiency. The TX also integrates an on-chip Vivaldi antenna with a measured realized gain of $5.0~\pm ~1.0$ dBi across 220–280 GHz after proton irradiation. The four-element phased-array TX achieves a maximum data rate of 60 Gb/s using 16 quadrature modulation (QAM) over 6 cm and 56 Gb/s using quadrature phase shift keying (QPSK) over 20 cm, with a peak effective isotropic radiated power of 16.2 dBm at 245 GHz. A 2-D $4\times 4$ beam pattern was measured using a stacked printed circuit board (PCB) configuration, demonstrating scanning coverages of ±24° in the $E$ -plane and ±28° in the $H$ -plane.
Resource
2026 EN
Elissar Al Aawar · Sofien Resifi · Juan Felipe Mendez Espinosa
+1 more
Accurate building footprint delineation and height estimation are critical for urban mapping, infrastructure monitoring, and risk assessment. We present SwinH-Fuse , a hierarchical satellite remote sensing framework that integrates data from the European Space Agency's (ESA) Sentinel-1 synthetic aperture radar (SAR) mission and Sentinel-2 multi-spectral optical mission to jointly extract building footprints and estimate heights. The framework employs a dual-branch Swin-UNet transformer for city-scale classification, a UNet for neighborhood-scale refinement, log-scale regression, and bias correction for height estimation. Monte Carlo dropout is integrated to quantify predictive uncertainty. Across diverse test regions in Europe and North America, SwinH-Fuse achieved a height RMSE of 1.22 m with a Pearson correlation coefficient of 0.73 with reference values. In an out-of-sample case study on the King Abdullah University of Science and Technology (KAUST) campus, the regression module maintained strong correlation despite classification challenges in sparse desert conditions. The results demonstrate that SwinH-Fuse provides accurate, scalable, and uncertainty-aware urban mapping suitable for sustainable planning and risk assessment.
Resource
2026 EN
Haitian Wang · Hezam Albaqami · Xinyu Wang
+5 more
(LiDAR)-based 3-D mapping suffers from cumulative drift causing global misalignment, particularly in global navigation satellite system (GNSS)-constrained environments. To address this, we propose a unified framework that fuses LiDAR, GNSS, and inertial measurement unit (IMU) data for high-resolution city-scale mapping. The method performs velocity-based temporal alignment using dynamic time warping and refines GNSS and IMU signals via extended Kalman filtering. Local maps are built using normal distributions transform-based registration and pose graph optimization with loop closure detection, while global consistency is enforced using GNSS-constrained anchors followed by fine registration of overlapping segments. We also introduce a large-scale multimodal dataset captured in Perth, Western Australia to facilitate future research in this direction. Our dataset comprises 144 000 frames acquired with a 128-channel Ouster LiDAR, synchronized RTK–GNSS trajectories, and MEMS–IMU measurements across 21 urban loops. To assess geometric consistency, we evaluated our method using alignment metrics based on road centerlines and intersections to capture both global and local accuracy. The proposed framework reduces the average global alignment error from 3.32 to 1.24 m, achieving a 61.4% improvement, and significantly decreases the intersection centroid offset from 13.22 to 2.01 m, corresponding to an 84.8% enhancement. The constructed high-fidelity map and raw dataset are publicly available through IEEE Dataport and its visualization can be viewed in the provided Demo. This dataset and method together establish a new benchmark for evaluating 3D city mapping in GNSS-constrained environments, with source code available at GitHub Repository.