Accurate information is a prerequisite for circulation and greater longevity of materials and products, enabling the economically and environmentally successful transition of circular economy (CE) in the aspiration towards sustainable development. However, a lack of appropriate digital tools is noticed in supplying relevant information and supporting recyclability, reusability, and reducibility processes. This study focuses on a building block of evolving technologies anointed the digital twin (DT). It aims to conceptualize DT as a cornerstone for the CE paradigm, positively influencing sustainability. It also endeavors to scrutinize remarkable implementations of DT-driven CE, accomplishing sustainable development goals (SDGs) in Ireland. An iterative methodology based on four steps is adopted to review an explosion of relevant, recent literature published in 2019–2024 and multiple case studies of Irish construction and manufacturing sectors. By analyzing the literature, this study discovers the evolution and taxonomy of DT, other technologies essential for its deployment, its proactive role in implementing CE and SDGs, and some challenges. Key findings of case studies reveal the current trajectory of employing DT with other digital products, upholding the CE practices within the bounds of the SDGs in Irish companies. In conclusion, DT is a conceivable means for advancing CE solutions and attaining SDGs towards social, economic, and environmental sustainability. The implications of this study are both theoretically and practically significant, encouraging industrialists and academia to adopt DT. It will help scholars gain a thorough understanding and guide industrialists to downsize their carbon footprint and waste, accomplishing SDGs. It will inspire managers with DT-driven information for decision-making and facilitate government and policymakers to collaborate in designing essential policies. Graphical Abstract

Urban smart city traffic management increasingly relies on UAV-based sensing, yet many widely used drone datasets annotate vehicles with axis-aligned bounding boxes that include unnecessary background and do not encode vehicle orientation. We present UAV-OBB, an aerial urban vehicle dataset with oriented bounding boxes (OBBs), designed for rotation-aware computer vision object detection and traffic monitoring from predominantly nadir-view UAV imagery. UAV-OBB contains 1617 RGB images at 1920 × 1080 resolution captured over roads in Chongqing and Wuhan (China), together with OBB Nannotations in YOLOv8-OBB label format and supplementary MP4 evaluation videos. The dataset provides 46,807 oriented annotations across six vehicle classes: bike, bus, car, other_vehicle, taxi, and truck, and is split into 1383 training images, 218 validation images, and 16 test images. Data were collected at 75 to 108 m altitude under diverse real-world conditions, including morning, midday, evening, night, rain, and mist or light fog, with both wide field-of-view and zoom settings to introduce strong scale variation. All instances were manually annotated using rotation-capable tools and double-checked for consistency, and occluded and truncated vehicles were included when the majority of the object was visible. To support practical smart city evaluation beyond static mAP, UAV-OBB also includes a short video clip with sparsely annotated reference frames and a longer unannotated sequence for qualitative assessment of temporal stability and deployment behaviour. UAV-OBB provides a realistic benchmark for rotation-aware detection, tracking, counting, and traffic flow analysis in urban UAV surveillance scenarios.

We are in a transformative era, and advances in Artificial Intelligence (AI), especially the foundational models, are constantly in the news. AI has been an integral part of many applications that rely on automation for service delivery, and one of them is mission-critical public safety applications. The problem with AI-oriented mission-critical applications is the humanin-the-loop system and the lack of adaptability to dynamic conditions while maintaining situational awareness. Agentic AI (AAI) has gained a lot of attention recently due to its ability to analyze textual data through a contextual lens while quickly adapting to conditions. In this context, this paper proposes an AAI framework for mission-critical applications. We propose a novel framework with a multi-layer architecture to realize the AAI. We also present a detailed implementation of AAI layer that bridges the gap between network infrastructure and missioncritical applications. Our preliminary analysis shows that the AAI reduces initial response time by 5.6 minutes on average, while alert generation time is reduced by 15.6 seconds on average and resource allocation is improved by up to 13.4%. We also show that the AAI methods improve the number of concurrent operations by 40, which reduces the recovery time by up to 5.2 minutes. Finally, we highlight some of the issues and challenges that need to be considered when implementing AAI frameworks.

Large Multimodal Models (LMMs) have achieved remarkable performance across vision-language tasks, yet their robustness against adversarial attacks remains critically underexplored. While LMMs are vulnerable to visual encoder attacks, they exhibit surprising resilience due to encoder diversity—attacks optimized for CLIP fail to transfer to EVA-CLIP, especially when textual context is provided. We introduce the Adaptive Ensemble PGD (AE-PGD) attack, which simultaneously targets both encoders through three key innovations: (1) dynamic adversarial caption selection , combining gradient magnitude with global semantic displacement to identify the most attack-effective caption per model; (2) an adaptive weight controller , dynamically balancing each encoder’s contribution using real-time loss, gradient norm, and confidence metrics; and (3) an Expectation over Transforms (EoT) gradient update ensuring robustness against input-transformation defenses. Evaluated on COCO 2014 images, AE-PGD reduces accuracy from a 75.42% baseline to 0.0% across all three evaluation metrics—visual encoding, image-to-text recall, and LLM answer recall—achieving complete model collapse. Manifold analysis confirms that adversarial perturbations push image embeddings to antipodal regions of the joint embedding space, activating semantically opposite concept clusters and producing structured hallucinations. WordNet WUP similarity analysis reveals a 33.5 percentage point semantic drop across the test set. AE-PGD causes state-of-the-art LMMs (LLaVA, Qwen-VL, GPT-4V) to catastrophically misidentify a bullet train as a “helicopter crash,” with strong black-box transfer yielding a 65 percentage point recall collapse on unseen encoders. This work exposes critical vulnerabilities in current LMM architectures and underscores the urgent need for ensemble-aware defense mechanisms.