Rapid urbanization in the twenty-first century has significantly accelerated the adoption of artificial intelligence (AI) technologies to address growing challenges in governance, mobility, energy, and urban security. This paper explores how AI is transforming smart city infrastructure, analyzing more than 92 academic publications published between 2012 and 2024. Key AI applications ranging from predictive analytics in e-governance to machine learning models in renewable energy management and autonomous mobility systems are synthesized domain-wise throughout this study. This paper highlights the benefits of AI-enabled decision making, finds current implementation barriers, and discusses the associated ethical implications. Furthermore, it presents a research agenda that stresses data interoperability, transparency, and human–AI collaboration to steer upcoming advancements in smart urban ecosystems.

Industrial cyber-physical systems face escalating vulnerabilities from operational anomalies and cyberattacks threatening critical infrastructure. This article presents an integrated framework combining deep learning-based anomaly detection, multiclass event classification, and authenticated encryption for real-time PMU data streams. A BiLSTM-transformer autoencoder achieves 97.79% anomaly detection accuracy with 62% fewer false positives and 90% fewer false negatives than baseline methods such as (e.g., simple autoencoder, LSTM). A BiLSTM-transformer classifier distinguishes six event types normal operation, equipment failures, line outages, generation outages, load changes, and cyber-induced bad data—with 98.27% accuracy across all classes. ChaCha20-HMAC-SHA256 authenticated encryption successfully mitigates 100% of simulated man-in-the-middle (MITM) and replay attacks, demonstrating 8.00 bits entropy, near-zero correlation ($\rho \approx 0$), and statistical indistinguishability. The NVIDIA Jetson Orin Nano validate deployment feasibility also facilitates GPU-accelerated multistream processing for substation-scale implementations. Evaluation of adversarial robustness indicates an accuracy of 86% under fast gradient sign method and projected gradient descent assaults, with adversarial training diminishing susceptibility by 63.5% . The architecture exhibits scalability and practical relevance for industrial cyber-physical security.

To ensure the reliability and sustained performance of industrial photovoltaic (PV) systems, fault detection frameworks must achieve both high detection accuracy and computational efficiency, particularly for deployment on resource-constrained edge platforms. This work proposes a lightweight and low-latency photovoltaic defect detection framework that integrates DCGAN-based generative augmentation with the proposed GhostViT-YOLOv10n architecture. The augmentation strategy helps address class imbalance, improve representation of rare defects, and enhance generalization capability in electroluminescence (EL) imagery through structured geometric and photometric transformations. The proposed framework integrates lightweight Ghost-based optimization, Cross-Stage Partial Fusion (C2f), Spatial Pyramid Pooling—Fast (SPPF), MobileViT contextual learning, and SimAM-based attention refinement to improve multi-scale feature extraction while maintaining low computational complexity. Experimental evaluation on the PVEL-AD and PV Multi Defect benchmark datasets demonstrates strong detection performance. On the PVEL-AD dataset, the BaseLine achieves a mAP@0.5 of 71.6% with only 2.7 M parameters and 8.4 GFLOPs, while our proposed GhostViT-YOLOv10n framework with DCGAN-enhanced version further improves detection performance to 93.6% mAP@0.5 with only 2.19 M parameters and 6.6 GFLOPs. On the PV Multi Defect dataset, the BaseLine achieves a mAP@0.5 of 74.0% with 2.71 M parameters and 8.4 GFLOPs, and the optimized framework with DCGAN-augmented configuration further improves performance to 95.4% mAP@0.5 with 2.58 M parameters and 7.7 GFLOPs. These results demonstrate the effectiveness of combining lightweight architectural optimization with generative augmentation for improving rare defect representation and multi-scale photovoltaic defect detection. To validate practical deployment feasibility, the optimized framework was deployed on a Raspberry Pi 5 using ONNX Runtime under CPU-only conditions. The deployed model achieved an average inference time of 43.05 ms and a real-time processing speed of 23.23 FPS while maintaining moderate CPU utilization and stable thermal behavior. These deployment results demonstrate the suitability of the proposed framework for lightweight edge-oriented photovoltaic inspection applications without requiring GPU acceleration. All evaluations were conducted exclusively on real test datasets, while synthetic samples were used only during training to improve data diversity and rare defect representation. Overall, the proposed framework provides a balanced solution that combines detection accuracy, computational efficiency, lightweight edge deployment capability, and generative augmentation for practical photovoltaic defect inspection applications with potential suitability for future drone-assisted inspection scenarios.

Chronic stress and depression-adjacent emotional states affect over 970 million people worldwide, yet their continuous, objective detection from physiological signals remains unsolved, particularly when cues are subtle and conventional approaches fail. Single modality classifiers capture only part of the picture, and binary valence/arousal formulations collapse emotionally distinct states into the same category, leaving conditions like sadness and depression characterised by Low-Valence Low-Arousal (LVLA) responses without reliable detection. A hybrid deep learning model combining Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM) networks, and Transformer encoders was developed to jointly classify four emotional quadrants from multimodal physiological data in the Database for Emotion Analysis using Physiological signals (DEAP) dataset. Electroencephalography (EEG), Galvanic Skin Response (GSR), Blood Volume Pulse (BVP), and respiration signals from 32 participants were preprocessed using fourth-order Butterworth filtering, trained with Adam optimisation, and evaluated through an 80/20 stratified split with five-fold cross-validation. The system achieved 91.2% four-quadrant valence–arousal accuracy (95% CI: 89.1–93.3%), with LVLA recall reaching 91.3%, outperforming all partial-hybrid variants. These findings demonstrate that hierarchical, attention-based fusion of physiological modalities can reliably distinguish stress from depression-adjacent states, offering a practical pathway toward continuous, non-invasive mental health monitoring on wearable platforms.