Pub Date : 2022-03-28DOI: 10.1109/ciot53061.2022.9766553
Anniek Eerdekens, Arne Callaert, M. Deruyck, L. Martens, W. Joseph
Sensor-based behavioral detection and classification can improve dog health and welfare. Since continuous monitoring is required, an energy-efficient solution is needed. The number of logging axes, sampling rate, and selected features of accelerometer data not only have a significant impact on classification accuracy in activity recognition but also on the sensor's energy needs. Three models are designed for detecting dog's activities namely, a Random Forest classifier (RF), a Convolutional Neural Network (CNN) and a hybrid CNN, i.e. a CNN fused with statistical features that retain knowledge about the global time series form. The models are validated using an experimental dataset consisting of six different dogs performing in eight different activities i.e. lying, sitting, standing, walking, running, sprinting, eating and drinking. The results indicate that using neck and chest accelerometer data sampled at 10 Hz is sufficient for high overall classification accuracies (96.44%) for the three models. The hybrid CNN is capable of excellent performance, detecting nearly 97.87% of the behaviours at 10 Hz with a class accuracy of 80 % or higher.
{"title":"Dog's Behaviour Classification Based on Wearable Sensor Accelerometer Data","authors":"Anniek Eerdekens, Arne Callaert, M. Deruyck, L. Martens, W. Joseph","doi":"10.1109/ciot53061.2022.9766553","DOIUrl":"https://doi.org/10.1109/ciot53061.2022.9766553","url":null,"abstract":"Sensor-based behavioral detection and classification can improve dog health and welfare. Since continuous monitoring is required, an energy-efficient solution is needed. The number of logging axes, sampling rate, and selected features of accelerometer data not only have a significant impact on classification accuracy in activity recognition but also on the sensor's energy needs. Three models are designed for detecting dog's activities namely, a Random Forest classifier (RF), a Convolutional Neural Network (CNN) and a hybrid CNN, i.e. a CNN fused with statistical features that retain knowledge about the global time series form. The models are validated using an experimental dataset consisting of six different dogs performing in eight different activities i.e. lying, sitting, standing, walking, running, sprinting, eating and drinking. The results indicate that using neck and chest accelerometer data sampled at 10 Hz is sufficient for high overall classification accuracies (96.44%) for the three models. The hybrid CNN is capable of excellent performance, detecting nearly 97.87% of the behaviours at 10 Hz with a class accuracy of 80 % or higher.","PeriodicalId":180813,"journal":{"name":"2022 5th Conference on Cloud and Internet of Things (CIoT)","volume":"139 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127473244","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Cloud Computing has become a point of interest to institutions, companies, and individuals because of the many advantages it provides, and its low-cost infrastructure, in addition to the ease of managing it and accessing it remotely anywhere there is an Internet connection that it has become a great innovation in the information technology. With the rapid evolution of cloud computing, information security concerns have emerged that hinder the evolution of cloud computing and need a solution, as security has become the main challenge of cloud computing. This paper will focus on cloud computing security, challenges, issues, threats, and solutions.
{"title":"Cloud Computing Security and Challenges: Issues, Threats, and Solutions","authors":"Sadeem Hamad Alrasheed, Majid Aied Alhariri, Sulaiman Abdulaziz Adubaykhi, S. Khediri","doi":"10.1109/ciot53061.2022.9766571","DOIUrl":"https://doi.org/10.1109/ciot53061.2022.9766571","url":null,"abstract":"Cloud Computing has become a point of interest to institutions, companies, and individuals because of the many advantages it provides, and its low-cost infrastructure, in addition to the ease of managing it and accessing it remotely anywhere there is an Internet connection that it has become a great innovation in the information technology. With the rapid evolution of cloud computing, information security concerns have emerged that hinder the evolution of cloud computing and need a solution, as security has become the main challenge of cloud computing. This paper will focus on cloud computing security, challenges, issues, threats, and solutions.","PeriodicalId":180813,"journal":{"name":"2022 5th Conference on Cloud and Internet of Things (CIoT)","volume":"47 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130174210","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-28DOI: 10.1109/ciot53061.2022.9766613
Preechai Mekbungwan, G. Pau, K. Kanchanasut
In-network computation allows application functions to be computed within the network directly on raw sensor data, and publish real-time responses or alerts to users in the field. We propose to extend Named Data Networking (NDN) with in-network computation by embedding functions in an additional entity called Function Library, which is connected to the NDN forwarder in each NDN router. Function calls can be expressed as part of the Interest names with proper name prefixes for routing, with the results of the computation returned as NDN Data packets, creating an ActiveNDN network. Our main focus is on performing robust distributed computation, such as analysing and filtering raw data in real-time, as close as possible to sensors in an environment with intermittent Internet connectivity and resource-constrained computable IoT nodes. In this paper, we describe the design of ActiveNDN with a small prototype network as a proof of concept. Extensive simulation experiments were conducted to investigate the performance and effectiveness of ActiveNDN in large-scale wireless IoT networks. We also compared the real-time processing capabilities of ActiveNDN with those of centralised edge computing. It has been shown that with the proposed minimal changes to NDN, low latency can be achieved so that time-critical IoT data processing of sensor data can meet the required deadlines.
{"title":"In-network Computation for IoT Data Processing with ActiveNDN in Wireless Sensor Networks","authors":"Preechai Mekbungwan, G. Pau, K. Kanchanasut","doi":"10.1109/ciot53061.2022.9766613","DOIUrl":"https://doi.org/10.1109/ciot53061.2022.9766613","url":null,"abstract":"In-network computation allows application functions to be computed within the network directly on raw sensor data, and publish real-time responses or alerts to users in the field. We propose to extend Named Data Networking (NDN) with in-network computation by embedding functions in an additional entity called Function Library, which is connected to the NDN forwarder in each NDN router. Function calls can be expressed as part of the Interest names with proper name prefixes for routing, with the results of the computation returned as NDN Data packets, creating an ActiveNDN network. Our main focus is on performing robust distributed computation, such as analysing and filtering raw data in real-time, as close as possible to sensors in an environment with intermittent Internet connectivity and resource-constrained computable IoT nodes. In this paper, we describe the design of ActiveNDN with a small prototype network as a proof of concept. Extensive simulation experiments were conducted to investigate the performance and effectiveness of ActiveNDN in large-scale wireless IoT networks. We also compared the real-time processing capabilities of ActiveNDN with those of centralised edge computing. It has been shown that with the proposed minimal changes to NDN, low latency can be achieved so that time-critical IoT data processing of sensor data can meet the required deadlines.","PeriodicalId":180813,"journal":{"name":"2022 5th Conference on Cloud and Internet of Things (CIoT)","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130905595","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-28DOI: 10.1109/ciot53061.2022.9766540
Soumaya Bachtobji, Djamel Eddine Kouicem, Mouna Ben Mabrouk
Building Management Systems (BMS) is a set of software and hardware tools that enables the monitoring and the control of building's mechanical and electrical equipment. These systems take advantage of Internet of Things (IoT) and Information Technology (IT) domains to efficiently enhance the buildings management. Blockchain technology is a new emerging technology that has attracted many researchers this last years. Basically, it has many applications in the field of cryptocur-rency, finance and insurances fields. However, these last years, blockchain has known other applications beyond cryptocurrency domain, such as the IoT. In this paper, we shed the light on one specific application of blockchain technology to assist the design of efficient BMS. We propose an hierarchical architecture composed of IoT, fog and cloud layers where the blockchain is integrated on the different layers to enhance the security of the whole BMS systems. We show the main advantages of applying blockchain technology in BMS systems and how our architecture outperforms the existing architectures in terms of low latency, security and privacy.
{"title":"Towards Blockchain Based Architecture for Building Information Modelling (BIM)","authors":"Soumaya Bachtobji, Djamel Eddine Kouicem, Mouna Ben Mabrouk","doi":"10.1109/ciot53061.2022.9766540","DOIUrl":"https://doi.org/10.1109/ciot53061.2022.9766540","url":null,"abstract":"Building Management Systems (BMS) is a set of software and hardware tools that enables the monitoring and the control of building's mechanical and electrical equipment. These systems take advantage of Internet of Things (IoT) and Information Technology (IT) domains to efficiently enhance the buildings management. Blockchain technology is a new emerging technology that has attracted many researchers this last years. Basically, it has many applications in the field of cryptocur-rency, finance and insurances fields. However, these last years, blockchain has known other applications beyond cryptocurrency domain, such as the IoT. In this paper, we shed the light on one specific application of blockchain technology to assist the design of efficient BMS. We propose an hierarchical architecture composed of IoT, fog and cloud layers where the blockchain is integrated on the different layers to enhance the security of the whole BMS systems. We show the main advantages of applying blockchain technology in BMS systems and how our architecture outperforms the existing architectures in terms of low latency, security and privacy.","PeriodicalId":180813,"journal":{"name":"2022 5th Conference on Cloud and Internet of Things (CIoT)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122160534","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-28DOI: 10.1109/ciot53061.2022.9766759
Hind Khoulimi, M. Lahby, Othman Benammar
Nowadays, the security of information system has become more and more important in our lives. Indeed, the appearance of 5G see 6G and technological progress which has given rise to the democratization of connected objects, thus increasing the related risks and making the task of information system security administrator more and harder. To remedy this, the researchers focused on several systems including IDS which is an Intrusion Detection System used in host and network security. However, this system generates a large number of alarms which must be managed by a security administrator, something which is not easy to do, but is necessary to guarantee an optimal level of security. In this work, we will present a system that helps the security administrator to properly detect and manage IDS alerts. This system is based on detecting attacks, collecting alerts generated by different IDS in a network of objects, analyzing these alerts and taking appropriate actions. We propose automation of said tasks based on artificial intelligence algorithms, especially Deep Learning. Our choice is directed towards the algorithm of the Artificial Neural Network (ANN) according to several criteria namely the performance and the speed of detection which is our major concern while combining it with the algorithm of Spider Monkey Optimization (SMO) for a good optimization of the entries. Our system aims to strengthen the second line of defense and make it more efficient and intelligent by equipping it with three intelligent engines namely, a detection engine, an analysis engine and an action engine. To illustrate the applicability of the proposed approaches, we begun to test the performance of detection by using different measures for example error of detection, training time and accuracy rate which have been obtained by testing with NSL-KDD dataset.
{"title":"Towards an intelligent system to manage IDS for IoT","authors":"Hind Khoulimi, M. Lahby, Othman Benammar","doi":"10.1109/ciot53061.2022.9766759","DOIUrl":"https://doi.org/10.1109/ciot53061.2022.9766759","url":null,"abstract":"Nowadays, the security of information system has become more and more important in our lives. Indeed, the appearance of 5G see 6G and technological progress which has given rise to the democratization of connected objects, thus increasing the related risks and making the task of information system security administrator more and harder. To remedy this, the researchers focused on several systems including IDS which is an Intrusion Detection System used in host and network security. However, this system generates a large number of alarms which must be managed by a security administrator, something which is not easy to do, but is necessary to guarantee an optimal level of security. In this work, we will present a system that helps the security administrator to properly detect and manage IDS alerts. This system is based on detecting attacks, collecting alerts generated by different IDS in a network of objects, analyzing these alerts and taking appropriate actions. We propose automation of said tasks based on artificial intelligence algorithms, especially Deep Learning. Our choice is directed towards the algorithm of the Artificial Neural Network (ANN) according to several criteria namely the performance and the speed of detection which is our major concern while combining it with the algorithm of Spider Monkey Optimization (SMO) for a good optimization of the entries. Our system aims to strengthen the second line of defense and make it more efficient and intelligent by equipping it with three intelligent engines namely, a detection engine, an analysis engine and an action engine. To illustrate the applicability of the proposed approaches, we begun to test the performance of detection by using different measures for example error of detection, training time and accuracy rate which have been obtained by testing with NSL-KDD dataset.","PeriodicalId":180813,"journal":{"name":"2022 5th Conference on Cloud and Internet of Things (CIoT)","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132935941","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-28DOI: 10.1109/ciot53061.2022.9766670
Noor O. Ahmed
This paper introduces a highly scalable Micro-services based Publish and Subscribe (MPaS) system model for IoT. MPaS follows design concepts from Software Define Networking and Fractal Theory, a subset of Chaos Theory, to decompose the traditional pub/sub system into a control and data planes and further into self-similar fractals implemented as self-replicating micro services that are correct by construction. We first present the system model. Then, discuss the MPaS prototype implementation built on top of ZeroMQ, an open source messaging middleware, and deployed in docker containers on Raspberry Pi computing cluster. Finally, we show the preliminary experimental results to illustrate the practicality and the efficacy of the proposed system model.
{"title":"MPaS: A Micro-services based Publish/Subscribe Middleware System Model for IoT","authors":"Noor O. Ahmed","doi":"10.1109/ciot53061.2022.9766670","DOIUrl":"https://doi.org/10.1109/ciot53061.2022.9766670","url":null,"abstract":"This paper introduces a highly scalable Micro-services based Publish and Subscribe (MPaS) system model for IoT. MPaS follows design concepts from Software Define Networking and Fractal Theory, a subset of Chaos Theory, to decompose the traditional pub/sub system into a control and data planes and further into self-similar fractals implemented as self-replicating micro services that are correct by construction. We first present the system model. Then, discuss the MPaS prototype implementation built on top of ZeroMQ, an open source messaging middleware, and deployed in docker containers on Raspberry Pi computing cluster. Finally, we show the preliminary experimental results to illustrate the practicality and the efficacy of the proposed system model.","PeriodicalId":180813,"journal":{"name":"2022 5th Conference on Cloud and Internet of Things (CIoT)","volume":"77 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114246400","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-28DOI: 10.1109/ciot53061.2022.9766544
Mostafa Alawieh, W. Fahs, Jamal Haydar, Fadlallah Chbib, A. Fadlallah
Intelligent Transportation System (ITS) has witnessed some great advancements in the recent years increasing the drivers' safety and comfort. ITS offers for types of wireless communications: vehicle to vehicle (V2V), vehicle to Road-Side Unit (V2R), Vehicle to everything (V2X) and Vehicle to Infrastructure (V2I). The highly dynamic connections and the time sensitivity of ITS make it an eye-catching field for attackers. Moreover, Security dangers and vulnerabilities have increased as V2V communications have evolved. Vehicular communications are exposed to several vulnerabilities such as Denial of Service attacks (DOS), Black hole and fabrication attacks. Fabrication attacks involve a malicious vehicle that alters the packet's information, causing serious network damage such as congestion and high delay. We propose a new secure routing scheme aims at detecting malicious vehicles and eliminating them from the network. Each vehicle is monitored by its neighbors. We identify a malicious vehicle by detecting an asymmetric average between its inputs and outputs routing packets, and a delete message is sent throughout the network to eliminate the malicious car from the routing table. Simulations show that the proposed scheme has improved the performance and efficiency of the Ad-hoc On-demand Distance Vector (AODV) protocol with respect to end-to-end delay, throughput and packet delivery ratio.
{"title":"A Secure Scheme for Vehicle-to-Vehicle (V2V) Routing Protocol","authors":"Mostafa Alawieh, W. Fahs, Jamal Haydar, Fadlallah Chbib, A. Fadlallah","doi":"10.1109/ciot53061.2022.9766544","DOIUrl":"https://doi.org/10.1109/ciot53061.2022.9766544","url":null,"abstract":"Intelligent Transportation System (ITS) has witnessed some great advancements in the recent years increasing the drivers' safety and comfort. ITS offers for types of wireless communications: vehicle to vehicle (V2V), vehicle to Road-Side Unit (V2R), Vehicle to everything (V2X) and Vehicle to Infrastructure (V2I). The highly dynamic connections and the time sensitivity of ITS make it an eye-catching field for attackers. Moreover, Security dangers and vulnerabilities have increased as V2V communications have evolved. Vehicular communications are exposed to several vulnerabilities such as Denial of Service attacks (DOS), Black hole and fabrication attacks. Fabrication attacks involve a malicious vehicle that alters the packet's information, causing serious network damage such as congestion and high delay. We propose a new secure routing scheme aims at detecting malicious vehicles and eliminating them from the network. Each vehicle is monitored by its neighbors. We identify a malicious vehicle by detecting an asymmetric average between its inputs and outputs routing packets, and a delete message is sent throughout the network to eliminate the malicious car from the routing table. Simulations show that the proposed scheme has improved the performance and efficiency of the Ad-hoc On-demand Distance Vector (AODV) protocol with respect to end-to-end delay, throughput and packet delivery ratio.","PeriodicalId":180813,"journal":{"name":"2022 5th Conference on Cloud and Internet of Things (CIoT)","volume":"152 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123271407","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-28DOI: 10.1109/ciot53061.2022.9766801
Farah Chahlaoui, H. Dahmouni, H. E. Alami
Due to the rapid growth of traffic and the increasing demand of network quality of service (QoS) and high throughput in today's 5G networks, load-balancing mechanisms became the main strategy used to enhance network performance and user experience. The use of standard load-balancing methods, however, may hinder the network's ability to achieve QoS goals in terms of resources utilization and latency constraints. In this paper, we study multi-path load balancing in a centralized SDN-based network to achieve the best network performance possible and we propose a multi-path load-balancing mechanism that guarantees a better user experience through optimizing delay, jitter, and the packet-loss rate.
{"title":"Multipath-routing based load-balancing in SDN networks","authors":"Farah Chahlaoui, H. Dahmouni, H. E. Alami","doi":"10.1109/ciot53061.2022.9766801","DOIUrl":"https://doi.org/10.1109/ciot53061.2022.9766801","url":null,"abstract":"Due to the rapid growth of traffic and the increasing demand of network quality of service (QoS) and high throughput in today's 5G networks, load-balancing mechanisms became the main strategy used to enhance network performance and user experience. The use of standard load-balancing methods, however, may hinder the network's ability to achieve QoS goals in terms of resources utilization and latency constraints. In this paper, we study multi-path load balancing in a centralized SDN-based network to achieve the best network performance possible and we propose a multi-path load-balancing mechanism that guarantees a better user experience through optimizing delay, jitter, and the packet-loss rate.","PeriodicalId":180813,"journal":{"name":"2022 5th Conference on Cloud and Internet of Things (CIoT)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121994749","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-28DOI: 10.1109/ciot53061.2022.9766532
Meryeme Ayache, A. Gawanmeh, J. Al-Karaki
Due to current Covid-19 pandemic, several countries enforce lock-down to prevent pandemic outspread. Hence, the mode of delivery of health services shall change as physical visits are not allowed. As such, the need of tele-medicine service and remote diagnosis have become a necessity. To provide reliable, safe, secure, and sustainable tele-medicine consultancy services, the supporting IT infrastructure need to be transformed. Therefore, it is necessary to use new generation of information technologies such as loT, Blockchain, and cloud computing to transform the traditional medical systems to smart healthcare systems. In this paper, we propose a proof of concept (PoC) of an ameliorated version of our DASS-CARE framework that supports decentralized, accessible, scalable, and secure access to healthcare services based on Internet of Medical things (IoMT) and Artificial Intelligence (AI). In this paper, we propose DASS-CARE 2.0 that offers more medical services including: (a) the real time health monitoring, (b) the collaborative and secure access to medical records, (c) the storage of medical history diagnosis and prescriptions, and (d)the patient's discharge and bills' payments. The paper concludes with future changes to the framework that can furnish further services.
{"title":"DASS-CARE 2.0: Blockchain-Based Healthcare Framework for Collaborative Diagnosis in CIoMT Ecosystem","authors":"Meryeme Ayache, A. Gawanmeh, J. Al-Karaki","doi":"10.1109/ciot53061.2022.9766532","DOIUrl":"https://doi.org/10.1109/ciot53061.2022.9766532","url":null,"abstract":"Due to current Covid-19 pandemic, several countries enforce lock-down to prevent pandemic outspread. Hence, the mode of delivery of health services shall change as physical visits are not allowed. As such, the need of tele-medicine service and remote diagnosis have become a necessity. To provide reliable, safe, secure, and sustainable tele-medicine consultancy services, the supporting IT infrastructure need to be transformed. Therefore, it is necessary to use new generation of information technologies such as loT, Blockchain, and cloud computing to transform the traditional medical systems to smart healthcare systems. In this paper, we propose a proof of concept (PoC) of an ameliorated version of our DASS-CARE framework that supports decentralized, accessible, scalable, and secure access to healthcare services based on Internet of Medical things (IoMT) and Artificial Intelligence (AI). In this paper, we propose DASS-CARE 2.0 that offers more medical services including: (a) the real time health monitoring, (b) the collaborative and secure access to medical records, (c) the storage of medical history diagnosis and prescriptions, and (d)the patient's discharge and bills' payments. The paper concludes with future changes to the framework that can furnish further services.","PeriodicalId":180813,"journal":{"name":"2022 5th Conference on Cloud and Internet of Things (CIoT)","volume":"57 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126632041","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-28DOI: 10.1109/ciot53061.2022.9766636
Sofien Resifi, H. Hassan, K. Drira
Deep Learning (DL) models are very efficient for many applications including, computer vision, natural language processing… Yet DL models require important computation resources making it particularly difficult to deploy these applications in constrained environments such as the Internet of Things (IoT). Offloading DL models to the cloud is one solution to this problem but has a number of drawbacks related to the trade-off between efficiency and latency, and other privacy issues. In this paper we try to solve this problem using two approaches, first by sharing the DL model between the cloud and the device and second by optimising the execution of the model using early exiting where inputs do not need to execute the model entirely. Both approaches are optimized automatically in order to choose the best sharing point and the best exiting point according to input. The solutions proposed could be easily generalized and are independent of applications and offer a good alternative in order to execute DL models locally.
{"title":"Adapting Deep Learning models to IoT environments","authors":"Sofien Resifi, H. Hassan, K. Drira","doi":"10.1109/ciot53061.2022.9766636","DOIUrl":"https://doi.org/10.1109/ciot53061.2022.9766636","url":null,"abstract":"Deep Learning (DL) models are very efficient for many applications including, computer vision, natural language processing… Yet DL models require important computation resources making it particularly difficult to deploy these applications in constrained environments such as the Internet of Things (IoT). Offloading DL models to the cloud is one solution to this problem but has a number of drawbacks related to the trade-off between efficiency and latency, and other privacy issues. In this paper we try to solve this problem using two approaches, first by sharing the DL model between the cloud and the device and second by optimising the execution of the model using early exiting where inputs do not need to execute the model entirely. Both approaches are optimized automatically in order to choose the best sharing point and the best exiting point according to input. The solutions proposed could be easily generalized and are independent of applications and offer a good alternative in order to execute DL models locally.","PeriodicalId":180813,"journal":{"name":"2022 5th Conference on Cloud and Internet of Things (CIoT)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126068766","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：