Modern smart grids encounter several challenges, including the variability of renewable energy sources with power fluctuations of ±35% in solar farms across the Arabian Gulf countries, delays in centralized control, with latencies of 150–200 ms, including all time needed from sending data to receiving control commands, and increased susceptibility to cyber threats, with a 30% annual increase in grid intrusions. Incorporating renewable energy into smart grids presents significant challenges, including intermittent generation, characterized by ±30% solar volatility, slow fault detection, and response times exceeding 500 ms in traditional systems, as well as centralized system weaknesses. This paper presents SwarmNet-5G, a bio-inspired neural network that integrates stigmergic learning and spiking neural networks (SNNs) for decentralized grid management to resolve existing issues. It merges Internet of Things (IoT)-edge intelligence, spiking neuromorphic computing, and swarm-based stigmergy to enable grid optimization. Its design decreases latency by 40% compared to typical deep reinforcement learning (DRL). It achieves a fault detection accuracy of 99.62% within 8.17 ms and requires 50 ms for repair, aligning with Saudi Arabia's Vision 2030 goal of achieving 50% renewable energy. Simulation results performed on the IEEE 39-bus using a dataset show that the solution allocated power during peak demand, such as a 20% load increase, with an efficiency of 92%, while typical DRL reached 78%. Additionally, the model projected annual savings of.2B in maintenance costs and a reduction of 4.8 megatons in CO₂ emissions per year.

Autonomous vehicles (AVs), particularly self-driving cars, have produced a large amount of interest in artificial intelligence (AI), intelligent transportation, and computer vision. Tracing and detecting numerous targets in real-time, mainly in city arrangements in adversarial environmental conditions, has become a significant challenge for AVs. The effectiveness of vehicle detection has been measured as a crucial stage in intelligent visual surveillance or traffic monitoring. After developing driver assistance and AV methods, adversarial weather conditions have become an essential problem. Nowadays, deep learning (DL) and machine learning (ML) models are critical to enhancing object detection in AVs, particularly in adversarial weather conditions. However, according to statistical learning, conventional AI is fundamental, facing restrictions due to manual feature engineering and restricted flexibility in adaptive environments. This study presents the explainable artificial intelligence with fusion-based transfer learning on adverse weather conditions detection for autonomous vehicles (XAIFTL-AWCDAV) method. The XAIFTL-AWCDAV model's main aim is to detect and classify weather conditions for AVs in challenging scenarios. In the preprocessing stage, the XAIFTL-AWCDAV model utilizes a non-local mean filtering (NLM) method for noise reduction. Besides, the XAIFTL-AWCDAV model performs feature extraction by fusing three models: EfficientNet, SqueezeNet, and MobileNetv2. The denoising autoencoder (DAE) technique is employed to classify adverse weather conditions. Next, the DAE method's hyperparameter selection uses the Levy sooty tern optimization (LSTO) approach. Finally, to ensure the transparency of the model's predictions, XAIFTL-AWCDAV integrates explainable AI (XAI) techniques, utilizing SHAP to visualize and interpret each feature's impact on the model's decision-making process. The efficiency of the XAIFTL-AWCDAV method is validated by comprehensive studies using a benchmark dataset. Numerical results show that the XAIFTL-AWCDAV method obtained a superior value of 98.90% over recent techniques.

Unmanned aerial vehicles (UAVs) are being increasingly adopted as flexible remote sensing platforms for smart logistics applications, including warehouse inventory, last-mile delivery supervision, traffic flow analysis, port operations, and infrastructure inspection. Despite their advantages, reliable object detection in UAV-based remote sensing imagery remains challenging due to small object sizes, dense object distributions, arbitrary orientations, and complex backgrounds commonly encountered in logistics environments. Although recent YOLO-based detectors have shown promising performance, their effectiveness is often limited in high-resolution aerial scenes and under practical computational constraints imposed by UAV platforms. To address these challenges, this paper proposes DV-YOLO, an enhanced deep learning framework tailored for object detection in UAV-based remote sensing imagery for logistics-oriented applications. The proposed model extends YOLOv9 through a deeper and wider backbone architecture coupled with optimized feature fusion strategies that jointly exploit spatial and semantic representations. A novel cross-path fusion network at deep feature map (CPFNDFM) is introduced to improve the detection of small and densely distributed logistics-related objects such as vehicles, containers, and infrastructure elements. In addition, a lightweight connection aggregation (CA) module, inspired by VoVNet and ShuffleNetV2, is integrated to enhance feature reuse while maintaining computational efficiency suitable for real-time UAV deployment. Furthermore, a challenging benchmark dataset, termed harder vision drone, is constructed by combining and refining samples from VisDrone and DOTA to better reflect real-world UAV remote sensing scenarios in logistics environments. Extensive experimental evaluations conducted on VisDrone 2021, DOTA v2, and the proposed dataset demonstrate that DV-YOLO consistently outperforms state-of-the-art detectors, achieving up to 3.5% improvement in mean average precision (mAP) compared with YOLOv9. These results highlight the potential of the proposed framework to support robust, accurate, and efficient aerial perception for smart logistics and UAV-based remote sensing applications.

The integration of Unmanned Aerial Vehicles (UAVs) into Intelligent Transportation Systems (ITS) holds transformative potential for real-time traffic monitoring, a critical component of emerging smart city infrastructure. UAVs offer unique advantages over stationary traffic cameras, including greater flexibility in monitoring large and dynamic urban areas. However, detecting small, densely packed vehicles in UAV imagery remains a significant challenge due to occlusion, variations in lighting, and the complexity of urban landscapes. Conventional models often struggle with these issues, leading to inaccurate detections and reduced performance in practical applications. To address these challenges, this paper introduces CFEMNet, an advanced deep learning model specifically designed for high-precision vehicle detection in complex urban environments. CFEMNet is built on the High-Resolution Network (HRNet) architecture and integrates a Context-aware Feature Extraction Module (CFEM), which combines multi-scale feature learning with a novel Self-Attention and Convolution layer setup within a Multi-scale Feature Block (MFB). This combination allows CFEMNet to accurately capture fine-grained details across varying scales, crucial for detecting small or partially occluded vehicles. Furthermore, the model incorporates an Equivalent Feed-Forward Network (EFFN) Block to ensure robust extraction of both spatial and semantic features, enhancing its ability to distinguish vehicles from similar objects. To optimize computational efficiency, CFEMNet employs a local window adaptation of Multi-head Self-Attention (MSA), which reduces memory overhead without sacrificing detection accuracy. Extensive experimental evaluations on the UAVDT and VisDrone-DET2018 datasets confirm CFEMNet’s superior performance in vehicle detection compared to existing models. This new architecture establishes CFEMNet as a benchmark for UAV-enabled traffic management, offering enhanced precision, reduced computational demands, and scalability for deployment in smart city applications. The advancements presented in CFEMNet contribute significantly to the evolution of smart city technologies, providing a foundation for intelligent and responsive traffic management systems that can adapt to the dynamic demands of urban environments.

One essential component of the futuristic way of living in "smart cities" is the installation of surveillance cameras. There are a wide variety of applications for surveillance cameras, including but not limited to: investigating and preventing crimes, identifying sick individuals (coronavirus), locating missing persons, and many more. In this research, we provided a system for smart city outdoor item recognition using visual data collected by security cameras. The object identification model used by the proposed outdoor system was an enhanced version of RetinaNet. A state of the art object identification model, RetinaNet boasts lightning-fast processing and pinpoint accuracy. Its primary purpose was to rectify the focal loss-based training dataset's inherent class imbalance. To make the RetinaNet better at identifying tiny objects, we increased its receptive field with custom-made convolution blocks. In addition, we adjusted the number of anchors by decreasing their scale and increasing their ratio. Using a mix of open-source datasets including BDD100K, MS COCO, and Pascal Vocab, the suggested outdoor object identification system was trained and tested. While maintaining real-time operation, the suggested system's performance has been markedly enhanced in terms of accuracy.

Enhancing road safety globally is imperative, especially given the significant portion of traffic-related fatalities attributed to motorcycle accidents resulting from non-compliance with helmet regulations. Acknowledging the critical role of helmets in rider protection, this paper presents an innovative approach to helmet violation detection using deep learning methodologies. The primary innovation involves the adaptation of the PerspectiveNet architecture, transitioning from the original Res2Net to the more efficient EfficientNet v2 backbone, aimed at bolstering detection capabilities. Through rigorous optimization techniques and extensive experimentation utilizing the India driving dataset (IDD) for training and validation, the system demonstrates exceptional performance, achieving an impressive detection accuracy of 95.2%, surpassing existing benchmarks. Furthermore, the optimized PerspectiveNet model showcases reduced computational complexity, marking a significant stride in real-time helmet violation detection for enhanced traffic management and road safety measures.