Jul 2026· 2026 6th International Conference on Inventive Computation and Information Technologies (ICICIT)· pp. 88-96· 0 citations· 14 references
Abstract
The Internet of Things (IoT) ecosystem has expanded and produced massive volumes of real-time streaming data, that must be processed quickly and in an efficient way. The latency problem, bandwidth congestion and bandwidth limitations in the traditional cloud-centric processing architecture are significant challenges, particularly in latency-sensitive systems such as smart healthcare, industrial automation and intelligent transportation systems. While edge computing reduces the amount of data that needs to be sent, edge devices are typically more limited in resources, which means they cannot run complex deep learning models. To address these limitations, this paper proposes a new Edge–Cloud Collaborative Deep Learning (ECC-DL) Framework to orchestrate different aspects of inference and learning between the edge nodes and cloud server. The lightweight convolutional neural networks (CNNs) are deployed at the edge nodes to extract features and make initial inferences, and the heavier end models are deployed at the cloud end for global analysis and regular retraining. To optimize the partitioning of the workload between the resources, an adaptive workload partitioning algorithm is designed that considers latency, bandwidth and computational resources. By experimental evaluation, latency is reduced by 40% and bandwidth consumption is reduced by 30% and prediction accuracy is improved over the centralized approach.
The high rate of Internet of Things (IoT) deployments has contributed to high volumes of heterogeneous data that are generated under drastic conditions of energy and computing constraints. The paper is a fully energy-saving IoT data analytics framework viewed through the prism of Automated Machine Learning (AutoML). The described comparison between classical machine learning models (ML), deep learning models (DL), and hybrid models (using an ensemble) evaluates the models using metrics where the scores are multi-dimensional, i.e., they consist of such measures as accuracy, power consumption and energy efficiency, inference latency, CPU usage, and memory consumption. An experimental evaluation of a practical smart city dataset of IoT demonstrates that there is a pronounced trade-off between the predictive performance and power expenditure. Deep learning models are also power-intensive despite their accuracy being high. To surpass that, a hybrid ML-DL mixture composed of CNN, SVM, and Random Forest ones is proposed in the assistance of energy-related weighted averaging. The power consumption of the hybrid model is 28.18 W, and its accuracy is 0.897, which offers the trade-off of using the hybrid model in the resource-constrained and dynamic Internet of Things.
Vishal P. Jariwala, Nirali Shah, Pratikkumar A. Parmar· ITEGAM- Journal of Engineeri...· 0 citations
The rapid expansion of the Internet of Things (IoT) has resulted in massive volumes of data being generated by interconnected devices across various domains. Traditional cloud-centric architectures often struggle with issues such as high latency, bandwidth constraints, and data privacy risks when processing this data. Edge computing has emerged as an effective solution by enabling data processing closer to the data source, thereby improving response time and reducing network dependency. In recent years, Generative Artificial Intelligence (GenAI) has gained significant attention for its ability to generate insights, predictions, and adaptive responses from complex and dynamic datasets. This paper examines the integration of Generative AI with IoT and edge computing to enhance intelligent edge systems capable of real-time analytics and autonomous decision-making. It explores architectural frameworks, potential applications in areas such as smart cities, healthcare, industrial automation, and autonomous systems, as well as the advantages of improved efficiency, scalability, and privacy preservation. Additionally, the paper discusses the technical challenges associated with deploying generative models at the edge, including resource constraints, model optimization, security, and data management. Finally, it outlines future research directions aimed at developing scalable, secure, and energy-efficient GenAI-enabled edge computing ecosystems.
Mr. L. S. Shendge, Mrs. S. N. Patel, Ms. D. V. Sharma· International Journal of Lat...· 0 citations
The rapid spread of IoT-enabled surveillance systems has led to an exponential growth in video data, posing significant challenges in real-time processing, scalability, and intelligent decision-making. Traditional CCTV-based systems remain largely passive, relying on human monitoring and lacking the capability to extract meaningful insights from large-scale distributed data streams. In this paper, we propose ELISA (Edge-based Layered Intelligent Surveillance Architecture), a multi-layer deep learning framework for intelligent surveillance via edge-based big data analytics in IoT environments. ELISA is structured into six layers: data collection, network communication, edge processing, face detection, behavior detection, and decision-making. The edge layer performs data preprocessing and filtering close to the source, reducing latency and bandwidth consumption while enabling real-time analytics. The face detection module identifies individuals using deep feature embeddings, while the behavior detection module recognizes activities such as falling, hugging, loitering, and phone use through a deep learning-based classifier. The decision layer integrates these outputs to generate context-aware alerts and actionable insights. Experimental evaluation on diverse datasets for face recognition and behavior classification demonstrates the effectiveness of ELISA in terms of accuracy, scalability, and computational efficiency. The results confirm its ability to process large-scale distributed video data while maintaining near real-time performance. Overall, ELISA transforms conventional surveillance into an intelligent, edge-driven, and data-centric system suitable for next-generation IoT applications.
Sarah Moustafa, Louai Saker, M. Dbouk· Discover Computing· 0 citations
The Internet of Things (IoT) systems generate vast amounts of data from numerous connected devices, posing significant challenges in terms of data processing, latency, and bandwidth utilization. Traditional cloud-based analytics systems face limitations, especially in real-time data processing. Edge computing, which brings computation closer to the data source, presents an ideal solution to overcome these challenges. This paper explores the design and implementation of edge computing pipelines for distributed analytics in IoT systems. It discusses the architecture of edge computing in IoT, the role of distributed analytics, and the flow of data through edge analytics pipelines. The paper also examines the benefits of edge computing, such as reduced latency, bandwidth efficiency, and enhanced real-time decision-making. Furthermore, we address the challenges in scaling, securing, and maintaining edge devices and propose various applications across smart cities, industrial IoT, healthcare, and agriculture. Finally, the paper highlights emerging trends and future directions for enhancing edge analytics capabilities in the ever-evolving IoT ecosystem.
Jose Fernandez· International Journal of Dat...· 0 citations
The growing pace of the rise in the proportion of data-driven intelligent services in smart cities, industrial systems, and cyber-physical infrastructures have exacerbated the demand of an effective distributed deep-learning over cloud-fog computing activities. Although the cloud- fog architectures offer proximity and scalability requirements, effective learning is difficult because of the decentralization of data, heterogeneity of the system, and constraints of the communication. In this paper, a federated and collaborative deep learning optimization framework is introduced, which allows hierarchical learning of training models on the fog nodes under cloud coordination. It uses adaptive local training, weighted collaborative aggregation, and proximal regularization: it removes non-IID effects caused by data, and enables the stabilization of convergence. The fog nodes do decentralized optimization and the cloud dynamically assembles the models according to the data volume and resource conditions. The experimental findings prove that the given approach has a smaller final global loss of 0.043 than 0.061 and 0.058 of standard federated learning and hierarchical baselines. Moreover, communication cost is also lowered to 29 MB round, which is an improvement of about 3040. These findings justify the usefulness of the suggested framework in scalable and resilient deep learning on cloud-fog systems.
M. Shamila, Shatraboina Shashank, Bhaskar Vishwakarma et al.· 2026 International Conferenc...· 0 citations
The rapid proliferation of Internet of Things (IoT) devices has resulted in massive, continuous data generation, demanding scalable, low-latency, and energy-efficient processing methodologies. Traditional cloud-based machine learning (ML) inference pipelines often incur high energy consumption due to persistent server provisioning and inefficient resource utilization. This paper proposes an energy-efficient ML inference framework using serverless cloud functions that dynamically scale with IoT workloads. The architecture leverages event-driven execution, model optimization techniques (quantization, pruning, edge pre-filtering), and adaptive model selection based on workload intensity. Experimental evaluations conducted on widely used serverless platforms demonstrate significant reductions in energy consumption, cold-start latency, and operational cost while maintaining high inference accuracy. The study highlights the potential of serverless computing as a sustainable backbone for next-generation IoT–ML systems, offering guidelines for building carbon-aware and cost-efficient inference pipelines for real-world applications.
Silvia Diallo, Grace Ndlovu· International Journal of Art...· 0 citations