Short-term wind power forecasting is considered a critical challenge in smart grid management due to the nonlinear, unstable, and multi-scale noise characteristics of wind signals. Although recent advances in hybrid deep learning have improved the accuracy of short-term wind power forecasting, many state-of-the-art models usually consider signal denoising, residual decomposition, and attention mechanisms as independent modules without providing a unified solution. This paper proposes an end-to-end solution, W-HiTS-Attention (Wavelet Transform, N-stacked Hierarchical Interpolation for Time Series, Attention), which coherently integrates wavelet denoising, hierarchical residual learning from N-HiTS (Neural Hierarchical Interpolation), and an in-block self-attention mechanism. The proposed solution outperforms 21 benchmarks in accuracy, including state-of-the-art baselines such as N-BEATS, N-HiTS, TCN, Informer, Autoformer, LSTM, BiLSTM, GRU, and Prophet, achieving an RMSE of 55.56 W and an R2 of 0.9918. Furthermore, the results show that the proposed solution is efficient in terms of parameter count (0.033M), latency (0.0036 ms/sample), and training time, making it promising for low-latency inference in resource-constrained environments. The results show that the coherent integration of frequency preprocessing, hierarchical residual forecasting, and attention-based temporal refinement provides a robust, explainable, and deployable solution for short-term wind power forecasting.

While touch screen technology is largely available, people with intellectual and developmental disabilities might have difficulties using it due to erratic touch responses. This study included six adults with intellectual and developmental disabilities who were managing touch screen technology modified via an input adaptation module. Such a module allowed them to activate the screen and access leisure events and complete match-to-sample tasks with various response configurations (e.g., precise clicks, heavy/prolonged touch, and taps). The first question of the study was whether successful history with the modified technology would facilitate the participants’ use of standard touch screen technology. A second (subordinate) question was whether practice with standard touch screen technology would improve their use of it. To address these questions, the participants were presented with series of sessions with standard technology in alternation with series of sessions with the modified technology. The results showed that the participants were highly successful with the modified technology and partially (or minimally) successful with the standard technology. Only three of them seemed to improve their performance with the standard technology following practice. In conclusion, the modified touch screen technology was consistently effective in helping participants with intellectual and developmental disabilities who continued to have difficulties in using standard touch screen technology.

Titanium and its alloys are widely used in orthopedic and dental implantology for their corrosion resistance and biocompatibility supporting osseointegration; however, their usage is accompanied by release of wear debris that may induce inflammatory responses. The necessity of formation of multifunctional coatings that accelerate osseointegration and provide long-term mechanical stability of titanium implants remains highly relevant. We propose a new simple and scalable coating method based on the laser shock processing technique, with TiO2 and SrCO3 powder mix used as an absorption layer. Our results show that this treatment created an approximately 158.3 ± 35.8 μm thick coating consisting of a mixed SrTiO3-TiO2 phase. The hardness of this coating evaluated by Vickers microhardness measurements showed a hardness increase of 3.3 times compared to the initial titanium substrate. Piezoelectric force microscopy (PFM) analysis revealed the presence of a reverse piezoelectric effect in the obtained structure confirming the highly likely successful synthesis of coating impregnated with SrTiO3. This piezoelectric coating can be readily deposited onto titanium substrates using the proposed method, enabling exploration of potential biomedical applications in future research.

The global transition towards sustainable and resilient energy systems has emphasized the need for efficient utilization of renewable energy sources (RESs) and rapid electrification of transportation. However, smart grids must address the intermittency of solar and wind power while accommodating the growing demand from electric vehicles (EVs). Hence, in this paper, a data-driven energy management system (EMS) is proposed that combines multivariable forecasting, generation scheduling, and EV charging coordination in a dual-level decentralized framework to increase the efficiency, reliability, and scalability of modern power grids. First, short-term forecasts of solar irradiance, wind speed, and load demand are addressed via five machine learning models ranging from nonlinear to ensemble models. Accordingly, a unified CatBoost-based platform for forecasting these three variables is selected because of its better performance and accuracy. These forecasts are subsequently utilized in a mixed-integer linear programming (MILP) framework for optimal generation scheduling in the considered network, fulfilling load demand at reduced electricity and emission costs while maintaining grid stability. Finally, a priority-based scheme is proposed for charging/discharging coordination of the aggregated EVs, minimizing demand variability while fulfilling vehicles’ charging needs and maintaining their batteries’ lifetime. The superiority of the proposed method lies in integrating a multivariable forecasting pipeline, linear MILP generation scheduling, and battery-health-aware V2G coordination in a unified decoupled framework, unlike many recent frontier works that treat these capabilities independently. Simulation results, under different scenarios, confirm that the proposed intelligent EMS can significantly reduce operational fluctuations, satisfy load and EV demands, optimize RES utilization, and support system cost-effectiveness, sustainability, and resilience.

Artificial intelligence (AI) has reshaped various aspects of human lives, particularly through its capabilities to address complex sustainability challenges. Despite the rapid expansion of AI applications, their contribution to cleaner production and sustainable development remains underexplored, especially in developing nations. In Southeast Asia (SEA), where AI adoption has grown substantially across environmental, economic, and social dimensions, research that examines its role in cleaner production outcomes remains fragmented. In view of this gap, this study conducts a systematic literature review (SLR) of AI applications related to cleaner production and sustainable development by examining relevant themes, application areas, and sustainability dimensions addressed by AI, while evaluating the maturity of AI methodologies, alignment with cleaner production outcomes, and integration with circular economy and resource efficiency goals. Moreover, it investigates the barriers and challenges that constrain AI application and offers future research directions to advance AI deployment for cleaner production and sustainable development across SEA countries.

Respiratory diseases constitute a major global health burden, necessitating accurate and reliable diagnostic support tools. Conventional auscultation, despite its widespread clinical use, remains inherently subjective and susceptible to inter-observer variability. In this study, we propose a unified deep learning framework for the automated classification of respiratory sound recordings into four clinically relevant categories: Normal, Crackles, Wheezes, and Crackles + Wheezes. The experimental evaluation was conducted on a publicly available dataset comprising heterogeneous respiratory recordings collected from both patients with pulmonary pathologies and healthy individuals. All audio signals were subjected to standardized preprocessing procedures to enhance signal consistency and ensure reliable feature extraction across acquisition conditions. To ensure methodological rigor and prevent optimistic bias, a strict subject-independent validation strategy was adopted using 5-fold GroupKFold cross-validation based on patient identifiers. Six deep learning architectures were systematically implemented and comparatively evaluated under a controlled and reproducible training protocol, including convolutional (1D-CNN, Deep-CNN), recurrent hybrid (CNN–LSTM, CNN–BiLSTM), and attention-based (CNN–Attention, CNN–Transformer) models. Performance metrics were reported as mean ± standard deviation across folds. The CNN–Attention architecture achieved the best overall performance, yielding a Balanced Accuracy of 90.1% ± 1.8% and a macro F1-score of 89.7% ± 2.1%, demonstrating stable inter-patient generalization. These findings indicate that attention-enhanced hybrid architectures effectively capture both local spectral structures and long-range temporal dependencies inherent in respiratory signals. The proposed framework provides a robust foundation for subject-independent automated lung sound classification and contributes to the development of clinically reliable decision-support systems.

The work presented herein proposes DisasterReliefGPT, a multimodal AI system for automation in the areas of crisis communication and post-disaster assessment. The system integrates three tightly coupled components: a vision module called DisasterOCS for structural damage detection in satellite images, a Large Vision–Language Model (LVLM) for enhanced visual understanding and contextual reasoning, and a Large Language Model (LLM) to produce detailed, clear assessment reports. DisasterOCS relies on a ResNet34-based encoder with partial weight sharing and event-specific decoders, coupled with a custom MultiCrossEntropyDiceLoss function for multi-class segmentation on pre- and post-disaster image pairs. On the benchmark xBD dataset, the developed system reaches a high score of 78.8% in identifying F1-damage, making correct identifications of destroyed buildings with 81.3% precision, while undamaged structures are found with a very high value of 90.7%. From a combination of these components, emergency responders can immediately provide reliable and readable assessments of damage that can be used to directly support urgent decision-making.

In smart logistics and intelligent manufacturing scenarios, the deployment of Autonomous Guided Vehicles (AGVs) necessitates vision systems that balance stringent real-time constraints with high detection accuracy. However, contemporary lightweight models often struggle with multi-scale feature representation and precision degradation. To address these challenges, this study presents LSOD-YOLO, a tailored evolution of YOLO11n designed for embedded AGV systems. Our methodology focuses on three architectural innovations: (1) we propose a Lightweight Shared Convolution Detection (LSCD) head integrated with Group Normalization (GN) and a scale-adaptive mechanism to harmonize multi-scale feature responses; (2) we re-engineer the backbone using a Star-Net architecture enhanced by Gated MLPs and Depthwise Attention to refine local spatial modeling; and (3) we integrate multi-branch residuals and Channel Attention (CAA) into the C3k2-Star-CAA module to enhance robustness against occlusions and complex backgrounds. The experimental validation on a self-built AGV industrial dataset and COCO128 reveals a compelling performance leap: a 30 FPS increase in throughput and a 1.5% gain in precision, all achieved with 32.8% fewer parameters. These findings confirm that LSOD-YOLO achieves a superior trade-off between computational efficiency and reliability, showing great potential for seamless deployment in resource-constrained AGV visual tasks.

Large language models (LLMs) exhibit strong semantic reasoning capabilities for autonomous driving decision-making; however, their substantial inference latency poses a critical challenge for real-time closed-loop vehicle control. This study proposes an engineering-oriented framework to enable latency-constrained LLM-based decision-making by integrating bird’s-eye-view (BEV) structured perception with low-bit quantized inference. The BEV perception module compresses multi-view visual inputs into structured semantic representations, thereby reducing input redundancy and enhancing inference efficiency. In addition, 4-bit post-training quantization (PTQ), combined with an optimized inference engine, is employed to alleviate computational and memory bandwidth constraints during autoregressive decoding. Experiments conducted on the CARLA simulation platform under car-following, overtaking, and mixed driving scenarios—validated through 500 independent trials—demonstrate that the proposed framework substantially reduces end-to-end inference latency while maintaining stable decision-making performance. The results indicate that the system satisfies the 10 Hz real-time control requirement and significantly improves control quality, as evidenced by reduced collision rates and lower Average Jerk compared with both traditional imitation learning (Behavioral Cloning, BC) and the Transformer-based TransFuser baseline. Furthermore, sensitivity analyses confirm the robustness of the framework under environmental degradation and perception noise, underscoring the practical feasibility of deploying LLMs for safe and reliable closed-loop autonomous driving.

This study investigates whether inter-channel similarity, quantified using Pearson’s correlation, can be used as an indicator of electrode relevance in electroencephalography (EEG)-based motor imagery (MI) classification and compares this approach with a genetic algorithm (GA)-based electrode selection strategy. Electrode subsets were obtained using Pearson correlation ranking, a GA optimizing classification accuracy, and the reference-study electrode subset reported in prior work. All subsets were evaluated on the BCI Competition IV Dataset 2a using a unified classifier architecture, and the sensitivity to classifier hyperparameter configuration was analyzed. Pearson-based selection achieved accuracies of 75.8% (8 channels), 78.1% (10 channels), and 81.5% (12 channels), while the GA achieved 75.9% (8 channels), 78.8% (10 channels), and 80.0% (13 channels). The reference-study electrode subset reached 75.0% (8 channels) and 76.7% (10 channels). Although correlation-based selection yielded competitive performance, no consistent relationship was observed between inter-channel similarity and discriminative relevance, and classification performance showed notable sensitivity to hyperparameter settings. These findings indicate that inter-channel similarity alone is not sufficient to determine electrode importance in MI classification and support the use of data-driven, model-aware selection strategies for the design of efficient low-channel-count brain–computer interface systems.