In recent years, the tremendous growth of interconnected devices results in a new technology called Internet of Things (IoT). Cloud computing assists the IoT applications to store the data and perform computation in order to control and manage the vast amount of data generated by these IoT devices. But the major challenge in Cloud computing is to meet requirements of many real-time applications of Internet of Things. Whereas, Edge is a computing architecture that helps to communicate, manage, store, and processes the data that quickly returns the response. This is made possible by moving these functionalities closer to the end users. Edge computing and cloud computing are independent as wells as mutually beneficial to many applications. This paper discusses the overview of IoT, communication technologies and protocols required for IoT, data transfer in IoT. Cloud services to store, process and to analyze the data generated from IoT devices are explored.

Edge computing is a new paradigm enabling intelligent applications for the Internet of Things (IoT) using mobile, low-cost IoT devices embedded with data analytics. Due to the resource limitations of Internet of Things devices, it is essential to use these resources optimally. Therefore, intelligence needs to be applied through an efficient deep learning model to optimize resources like memory, power, and computational ability. In addition, intelligent edge computing is essential for real-time applications requiring end-to-end delay or response time within a few seconds. We propose decentralized heterogeneous edge clusters deployed with an optimized pre-trained yolov2 model. In our model, the weights have been pruned and then split into fused layers and distributed to edge devices for processing. Later the gateway device merges the partial results from each edge device to obtain the processed output. We deploy a convolutional neural network (CNN) on resource-constraint IoT devices to make them intelligent and realistic. Evaluation was done by deploying the proposed model on five IoT edge devices and a gateway device enabled with hardware accelerator. The evaluation of our proposed model shows significant improvement in terms of communication size and inference latency. Compared to DeepThings for 5 X 5 fused layer partitioning for five devices, our proposed model reduces communication size by ~ 14.4% and inference latency by ~16%.

Most real-time computer vision applications heavily rely on Convolutional Neural Network (CNN) based models, for image classification and recognition. Due to the computationally and memory-intensive nature of the CNN model, it’s challenging to deploy on resource-constrained Internet of Things (IoT) devices to enable Edge intelligence for real-time decision-making. Edge intelligence requires minimum inference latency, memory footprint, and energy-efficient model. This work aims to develop an energy-efficient deep learning accelerator using a 3-stage pipeline: Training, Weight-pruning, and Quantization to reduce the model size and optimize the resources. First, we employ YOLOv3, a CNN architecture to detect objects in an image on the trained data. In addition, a sparse network of YOLO has been created by using pruning, which helps to improve the network’s performance and efficiency by reducing the computational requirements. Finally, we utilize 8-bit quantization to reduce the precision of the weights and activations, in a neural network. The evaluation of our proposed model shows that combining pruning and 8-bit quantization improves the efficiency and performance of the model. While pruning shows a decline of 80.39% in model parameters. The combination of 8-bit quantization results in an improvement in inference latency by 22.72% compared to existing SQuantization approach and a reduction of energy consumption by 29.41%.

The ambulance dispatch system plays a crucial role in emergency medical care by ensuring efficient communication, reducing response times, and ultimately saving lives. Delays in ambulance arrival can have serious consequences for patient health and survival. To enhance emergency preparedness, decision trees are used to analyze historical data and predict ambulance demand in specific locations over time. This helps in planning the necessary number of ambulances in advance. In situations where ambulance resources are limited, a support vector machine (SVM) evaluates patient data to optimize the distribution of available ambulances, ensuring that the most critical patients receive timely medical attention. For real-time route optimization, a convolutional neural network (CNN)-based deep learning model is used to adjust ambulance routes based on current traffic and road conditions, achieving an accuracy of 99.15%. By improving dispatch efficiency and communication, the proposed machine learning-based ambulance system reduces the burden on emergency services, enhancing overall effectiveness, particularly during peak demand periods.

Person re-identification (Re-ID) emerges as a captivating realm within computer vision, dedicated to the task of recognizing the same individual across diverse camera angles or locations. The realm of video-based person re-identification (video re-ID) has recently captivated increasing interest, owing to its wide array of practical applications spanning surveillance, smart city solutions, and public safety measures. Nevertheless, video re-ID proves to be a formidable challenge, an ever-evolving domain fraught with a multitude of uncertainties like viewpoint variations, occlusions, pose changes, and unpredictable video sequences. Over the past few years, the realm of deep learning applied to video re-ID has consistently delivered remarkable outcomes on public datasets, showcasing a range of innovative strategies devised to tackle the array of issues encountered in video re-ID. In stark contrast to image-based re-ID, video re-ID stands out as significantly more intricate and demanding. In a bid to inspire forthcoming research endeavors and confronts emerging challenges, this paper presents a comprehensive overview of the latest advancements in deep learning methodologies tailored for video re-ID. It delves into three crucial facets; encompassing succinct explanations of video re-ID techniques along with their constraints, pivotal breakthroughs coupled with the technical hurdles faced, and the architectural framework underpinning these developments. The paper further furnishes a comparative analysis of performance across diverse datasets, offers insightful guidance on enhancing video re-ID strategies, and outlines compelling avenues for future research exploration.