The adoption of drones is rapidly expanding across various industries, including logistics, light shows, and agriculture. Beyond these applications, drones are also becoming essential tools for disaster response and emergency operations. This study focuses on the design of a fault-tolerant distributed control system for robust and efficient coordination of swarm drones. To enhance navigation efficiency and prevent collisions, a three-dimensional graph structure has been considered for drone positioning. This approach enables the implementation of more efficient pathfinding algorithms, while reducing the risk of mid-air conflicts. To overcome range limitations, the proposed system leverages LTE/4G connectivity for seamless communication between drones. Furthermore, key differences between centralized and distributed control architectures are discussed, highlighting their respective advantages and tradeoffs. A prototype system is developed for the experimentation of its behavior in a real-world scenario. The system is also tested in a simulation environment for further evaluation of its performance and reliability.

The proliferation of the Internet of Things (IoT) requires efficient and scalable data management solutions at the edge of the network. Edge devices play a critical role in processing data and making real-time decisions, thereby minimizing latency and reducing the load on central servers. However, these devices face challenges such as scalability, latency, and fault tolerance. This paper proposes a distributed key-value store specifically designed for IoT edge devices to address these challenges. By integrating distributed key-value stores with IoT edge devices, we aim to enhance data access and management efficiency. Our approach includes the development of optimal methodologies and architectural standards for this integration, addressing key issues such as data consistency, fault tolerance, and low latency. Through prototype creation and testing, we demonstrate significant improvements in data processing speed and system reliability. The ultimate objective of this research is to advance edge computing for IoT, fostering more robust, scalable, and efficient systems.

In the realm of Internet of Things (IoT) architectures where diversified services exist, latency related network challenges remain as critical problems in cloud based systems. To address these challenges, fog computing has emerged as a viable solution for Quality of Service (QoS) oriented design. Leveraging heterogeneous services with different requirements, fog layer transmits decision outputs with traffic patterns composed of multi-priority data packets. To fully harness the benefits of fog computing, it is crucial to model such data traffic and devise an efficient scheduling approach for its management. Additionally, employing effective management of cloud resources for handling incoming data traffic is pivotal to enhance the design of latency based QoS aware services. In this paper, we present a fog enabled service architecture and its novel IoT data traffic management mechanism for the mitigation of latency based QoS problems. To this end, we first elaborate on the service components and data traffic characteristics of the architecture with an in-depth analysis of arrival-service model to give insights into the starvation issue in multi-priority scheduling. We present a Complex Event Processing (CEP) based scheduling approach for handling multi-priority IoT data traffic with QoS perspective and a dynamic resource scaling mechanism for handling decision data traffic on the cloud. Finally, we discuss the test results obtained from implementation of the proposed policies and services in a real test-bed environment with two IoT use-cases. The results reveal that the proposed architecture significantly improves the wait time based QoS requirements and resource utilization compared to the baseline system.

Real-time vehicle detection is a critical component of traffic monitoring, with significant implications for smart city applications. Accurate and efficient detection of vehicles can improve traffic flow and reduce congestion. This paper presents a real-time vehicle detection system based on Deep Learning (DL) techniques, using the YOLOv7 object detection framework. The system was trained on a novel dataset with a diverse range of vehicles, including different sizes, orientations, and lighting conditions, to improve object detection accuracy. To reduce the required training time and computing resources, Transfer learning is utilized to fine-tune two variants of YOLOv7, YOLOv7-x and YOLOv7-tiny. The results of the experiment revealed that the YOLOv7-x achieved a mean Average Precision (mAP) rate of 96.7%, while the YOLOv7-tiny achieved 89.3%. Furthermore, when the models were fed with a video stream, the YOLOv7-tiny achieved 57 Frames Per Second (FPS), whereas the YOLOv7-x achieved 27 FPS. As a result, the YOLOv7-tiny is more suitable for resource-constrained devices, such as those frequently utilized in IoT applications, due to its smaller model size, lower computational requirements, and higher FPS rate with acceptable accuracy. On the other hand, if higher accuracy is the priority, the YOLOv7-x model should be considered. The proposed frameworks help to improve the effectiveness of traffic management systems, leading to more efficient and sustainable transportation in smart cities.

Studies on the Internet of Things have contributed to the development of air quality monitoring applications within the smart city paradigm. Furthermore, fog computing helps eliminate the data density bottleneck that may arise as smart city applications evolve. In this study, a fog computing-based, location-aware air quality monitoring system has been proposed, and its performance was evaluated through preliminary tests using real data. The proposed system is compared with a centralized cloud-based application scenario. The results indicate that the proposed system can serve up to 960 clients in the presence of 120 air quality monitoring stations in the established test environment. Additionally, the study revealed that the proposed model outperforms the cloud-based model in terms of latency and service load stability characteristics.

In this demonstration, we will construct and visualize a scalable fog computing-based multi-layer Internet of Things (IoT) platform that supports various application scenarios. Within the platform, we assess its performance in terms of latency, wait time, and processing load as QoS metrics. We examine scenarios where applications generating different alarm levels, resulting in multi-priority decision data traffic across clients and sensor nodes. This proposal provides an overview of the platform, including its layered design, the demo setup, and how QoS is addressed across the different layers.

Renewable energy is a hot-topic for many researchers and companies alike because of the environmental concerns about current means of energy production. However, many kinds of renewable energy sources generate fluctuating power, which is hard to load-balance through different days of a month, or even hours of a day. This consistency problem increases the need for energy storage improvement. In this study, we elaborate on if an IoT enabled smart-grid application can successfully manage energy consistency for households through using electric vehicles as mobile storage means. We designed and implemented a highly-realistic projection for a city of Turkey, Denizli to observe the impact of using electric vehicles within a solar powered grid. The simulation results reveal that the proposed solution provides self-sufficient uptimes more than twice of the standalone grid with uptimes up to 100%, showcasing an immense potential for such applications.

There is a growing need to transmit video data over IoT infrastructures in the same processing pipeline with sensor edge devices. As the number of devices in the same pipeline increases, consumers experience more frequent message delays while getting data from multiple sources. In this study, we propose and analyze three priority queue management algorithms for publish-subscribe based Kafka framework to utilize realtime video streaming. The algorithms are implemented on a real test-bed by deploying multiple sensors with different Quality of Service (QoS) requirements. The results reveal that the proposed solutions enhance QoS provisioning when compared to the default Kafka queue model. Moreover, the Updated Weighted Fair Kafka Queue (UWFKQ) algorithm performs better than the Weighted Fair Kafka Queue (WFKQ) and the Priority Kafka Queue algorithms for relieving starvation as well as long wait times. Besides, the slider mechanism proposed for UWFKQ algorithm enables the system to utilize unused poll counts so that congestion between packet transmission flows can be controlled. We have also shown that several default Kafka parameters can successfully be tuned to optimize the throughput rate.

With the proliferation of heterogeneous services in scalable Internet-of-Things (IoT) architectures, the Cloud of Things begins to face latency-based network challenges. As a solution to these problems, the Fog-of-Things (FoT) paradigm has emerged. This paradigm promises numerous benefits for latency-sensitive IoT service design, by reducing IoT data traffic toward the cloud. Since FoT utilizes heterogeneous services, the decision outputs of these services are transmitted by multipriority data packet traffic patterns. For the full utilization of fog computing benefits, we need to model this traffic and apply an efficient data traffic scheduling approach. In this study, we address the starvation challenge of multipriority FoT data traffic scheduling in a scalable fog architecture and propose a complex event processing-based efficient scheduling policy for its prevention. We first give the arrival–service model for FoT data traffic using finite-size multilevel waiting lines. Then, we compare the proposed policy with the first-in–first-out and multipriority-discipline queue policies through a comprehensive analysis of waiting times and wait-time gap characteristics. We have also conducted extensive simulation tests to explore the performance of these policies in our testbed for up to 800 clients communicating with the fog system. The results reveal that the empirical values obtained from the tests verify the theoretical model and the proposed approach can successfully relieve the waiting-time gaps observed in priority levels.

SummaryThe vast number of smart cloud applications that communicate with the “things” over a variety of physical networks and communication protocols contribute to the rise of complexity in Internet of Things (IoT) systems. The diversity of end‐user requirements related to the volume of generated data, its variety, and the velocity of its transmission makes quality of service (QoS) provisioning even more crucial and challenging in IoT. This paper provides a comprehensive and up‐to‐date survey of QoS support in IoT networks and communication protocols. An analysis of the QoS‐aware IoT architectures, layer‐dependent QoS metrics, and network resource optimization methods utilized in IoT systems are given. The limitations of the current state‐of‐the‐art studies for efficient delivery of QoS metrics are discussed. The paper concludes with future research directions on end‐to‐end QoS provisioning in IoT.

Internet of Things (IoT) services have grown substantially in recent years. Consequently, IoT service providers (SPs) are emerging in the market and competing to offer their services. Many IoT applications utilize these services in an integrated manner with different Quality-of-Service (QoS) requirements. Thus, the provisioning of end-to-end QoS is getting more indispensable for IoT platforms. However, provisioning the system by using only QoS metrics without considering user experiences is not sufficient. Recently, Quality of Experience (QoE) model has become a promising approach to quantify actual user experiences of services. A holistic design approach that considers constraints of various QoS/QoE metrics together is needed to satisfy requirements of these applications and services. Besides, IoT services may operate in environments with limited resources. Therefore, effective management of services and system resources is essential for QoS/QoE support. This paper provides a comprehensive survey for the state-of-the-art studies on IoT services with QoS/QoE perspective. Our contributions are threefold: 1) QoE-driven architecture is demonstrated by classifying vital components according to QoE-related functions in prior studies; 2) QoE metrics and QoE optimization objectives are classified by corresponding system and resource control problems in the architecture; and 3) QoE-aware resource management e.g., QoE-aware offloading, placement and data caching policies with recent Machine Learning approaches are extensively reviewed.

Fog computing-based IoT paradigm is gaining more importance as it solves high latency challenges of cloud computing-based IoT. Fog solutions promise numerous advantages to real-time IoT services in a heterogeneous network. However, it is essential to achieve effective resource provisioning and reliable data transfer inside the fog to utilize these advantages. In this study, we address the data transfer challenges in a scalable fog architecture and propose a message queue based fast reliable data transmission mechanism. We compare the proposed messaging model with a conventional subscribe-update-notify based data flow approach for the orchestration of reliable data transfer. We provide a comprehensive analysis of timing constraints and delay characteristics of both methods. We then present the performance results of extensive simulation scenarios for up to 100 clients communicating with 256 sensors simultaneously over the proposed platform. The results reveal that the message queue-based approach successfully relieves the latency cost of reliable data transmission in the inner fog.

Human activity recognition (HAR) enables numerous application scenarios in ambient assisted living thanks to IoT integration. Healthcare is one of the most prominent use cases of HAR serving individuals who suffer from aging or disabilities as well as healthy people. Remote monitoring of daily activities within home environment may offer assistance in tracking adherence of patients to therapeutical procedures such as exercise monitoring. Within an IoT framework, the type of sensors used influences usability of the HAR system. Accelerometers introduce a noninvasive sensing option as opposed to cameras which intrude into users’ privacy. Recognition of target activities when they are performed among other activities brings about challenges. Typical multi-class classification employed in such recognition tasks necessitates training data collection for all activity types which can be encountered in the prediction stage. Due to unlimited variety of daily living activities, the number of activity classes for which training data should be collected is infinitely many. Expressing the recognition problem in terms of one-class classification (OCC) architecture can aid in resolving this bottleneck. In this chapter, we propose an OCC-based HAR architecture with IoT integration. In our OCC scheme, we utilize artificial data generation (ADG) to generate training data for the negative class based on the target class. In the proposed model, the target class is the only class for which training data are collected. The OCC scheme enables recognizing the target class when the other class is represented with artificially generated data. We present the results of an experimental study for our OCC model on a dataset consisting of ambulatory and static activities.

Spotting the target activity in a sequence of activities and transitions without applying a predetermined window size is a challenging task. The sliding window method, which is the typical approach in segmenting the continuous data stream, operates with optimal segment sizes chosen considering the type and duration of the activities. Nevertheless, activity type and duration are usually unknown to the detection unit in practice. In this study, we proposed the nonpredetermined size windowing (NSW) scheme to spot the target activity performed in a sequence of unseen activities. NSW is built on classifying progress-based features in multilayer training and prediction stages where the time-domain progress is expressed in terms of polynomials. Thus, it operates without incorporating the information regarding duration and type of the activities. We verified our method with a proof-of-concept use case where data are acquired by a single wrist-worn 3-D accelerometer. We compared our method against fixed size windowing performed with varying window sizes and feature extraction schemes; windowed energy, peak frequency, Shannon entropy, and wavelet entropy. Our method outperforms the compared schemes, reaching a median accuracy of 89% and a median true positive rate of 1 in both intrasubject and intersubject cases based on an independent test set where each test instance contains a sequence of nine different activities.

Fog computing has the benefits to handle and reduce data traffic load towards the central cloud in IoT systems. These benefits are facilitated with the help of offloaded fog services that participate in the decision making processes. Besides, fog-based systems have the potential to mitigate scalability bottlenecks that occur in cloud-based systems. In this study, we elaborate on fog based design for a scalable real time air quality monitoring and alert generation system. We established an emulation test bed with real data collected from air quality sensing nodes deployed around Bangkok and vicinity areas to understand the behavior of the proposed solution in terms of waiting time characteristics. We analyzed the performance of the system in two design scenarios; first scenario is built with the proposed fog solution and the second one is the cloud-based approach. We present the performance results revealing the advantages of the proposed model, for the number of air box nodes scaling up to 120 and the number of client nodes up to 200.

In this study, we propose a smart plant irrigation IoT system that autonomously adapts itself to a defined irrigation habit. The automated plant irrigation systems generally make decisions based on static models derived from the plant's characteristics. In contrast, in our proposed solution, irrigation decisions are dynamically adjusted based on the changing environmental conditions. The learning mechanism of the model reveals the mathematical connections of the environmental variables used in the determination of the irrigation habit and progressively enhances its learning procedure as the irrigation data accumulates in the model. We evaluated the success of our irrigation model with four different supervised machine learning algorithms and adapted the Gradient Boosting Regression Trees(GBRT) method in our IoT solution. We established a test bed for the sensor edge, mobile client, and the decision service on the cloud to analyze the overall system performance. The early results from our prototype system that is tested with two indoor plants; namely Sardinia and Peace-lily are very encouraging. The results reveal that the proposed system can learn the irrigation habits of different plants successfully.

MQTT, AMQP and HTTP are messaging protocols that are commonly used for communicating with resource-constrained IoT devices. HTTP is the standard reference protocol for the REST transportation based on the request/response model, whereas both AMQP and MQTT are message-oriented protocols that use the publish/subscribe model. Message-oriented protocols enhance some of the shortcomings of the complex HTTP protocol by using asynchronous communication, changing the design from a document-centric to a data-centric approach and decreasing the header and message sizes. Although significant technical detail is present on these protocols, their real-time performance is insufficiently elaborated. In this paper, we present an experimental evaluation of these protocols conducted in a homogeneous IoT testbed using a real-time Smart City public data set. We provide the behavioral differences between messaging-based protocols and the REST-based HTTP protocol in terms of message latency and CPU usage for varying traffic loads and message sizes. The results showed that MQTT and AMQP are four times faster than HTTP protocol when comparing the message sent latencies. HTTP uses four times more CPU than the AMQP and MQTT protocols. In summary, message-oriented protocols give more stable and improved results as compared to the REST model-based HTTP protocol for all evaluation scenarios.

The main goal of the Internet of Things (IoT) platforms is to create different kinds of smart cloud data services that can respond in real-time to large amounts of sensor data received across these links. Although it has been more than twenty years since the first IoT networks were used, there are still several performance issues that need to be mitigated. This is a challenging issue since IoT scenarios have a wide range of requirements from the network. This led researchers to design different application layer messaging protocols to satisfy these requirements. While several studies investigate the messaging protocols in terms of performance, reliability, security, and energy consumption, almost none of them consider mobility in their comparisons. In this study, we analyze the effect of mobility on the performance of IoT networks. We established a simulation testbed using Netsim Simulator and conducted a set of experiments to observe the effect of mobility on the application layer protocols; HTTP and CoAP. Our results show that mobility can have a significant effect on the throughput, delay and battery consumption of the messaging protocols under consideration.

In this study, we propose a human activity recognition (HAR) system for monitoring the adherence of the patients to the elbow flexion and extension physiotherapy exercise routines in their daily environments. The proposed solution utilizes single wrist-worn accelerometer data with a one-class classification (OCC) approach where training data are collected for the target class only and data that represent all other classes are produced artificially from the target activity data. We analyzed the system with four different classifiers; KNN (k-nearest-neighbors), SVM (Support Vector Machines), logistic regression and Naive Bayes classifiers. The results reveal that SVM achieves a median success rate of 100% and 83,3% in detection of target and other classes respectively in intra-subject evaluation as well as reaching a median success rate of 90% and 100% in inter-subject case on a per-subject basis, emerging as the best performing classifier among the ones we study. Our one-class approach is also observed to outperform one-class SVM of LIBSVM, which severely suffers from parameter selection issue.

In this work, a rule-based data management system is proposed for heterogenous cloud data service architectures. For the management of data, many web services and an overarching control mechanism that communicates with these services quickly and accurately to obtain a reliable system are required. Hence, the proposed rule-based system communicates with the web services regularly and processes the designated rules using a multi-layer queuing model for scheduling. The data service architecture consists of three main parts: user management, rule management, and rule processing. These parts contain data services that require high security, the system safety is provided with a two-level protection mechanism namely authentication and authorization. Scheduling system consists of queues where jobs are inserted to be processed. The proposed system also considers the interdependence of sub-jobs. The queue system is created to use a priority queue. A custom Graduate School Administrative Board Decisions Management System (GSBS) is designed and implemented to test and verify the limits of the proposed system. The proposed multi-layer queuing approach is then compared with the commonly used FIFO approach. Test results demonstrate that the proposed Rule Based Distributed Data-Service (RBDDS) is efficient in terms of scalability, response time and throughput metrics.