Pub Date : 2022-05-01DOI: 10.1109/cniot55862.2022.00041
Zhan Shi, Lei-Lei Han, Ying Qian
The UML (Unified Modeling Language) language uses the graphical representation for a system, which has been widely used as an industrial standard in software engineering. However, many UML concepts are based on informal semantics, it is difficult to achieve accurate definitions of target models. On the basis of the existing TEGG grammar, this paper extends the TEGG to construct a formal method for UML state diagrams, and proposes productions for the transformation from UML state diagrams to TEGG graphs. Finally, a case is shown to illustrate the implementation of the proposed method.
{"title":"A Formalization and Transformation Method of UML Model","authors":"Zhan Shi, Lei-Lei Han, Ying Qian","doi":"10.1109/cniot55862.2022.00041","DOIUrl":"https://doi.org/10.1109/cniot55862.2022.00041","url":null,"abstract":"The UML (Unified Modeling Language) language uses the graphical representation for a system, which has been widely used as an industrial standard in software engineering. However, many UML concepts are based on informal semantics, it is difficult to achieve accurate definitions of target models. On the basis of the existing TEGG grammar, this paper extends the TEGG to construct a formal method for UML state diagrams, and proposes productions for the transformation from UML state diagrams to TEGG graphs. Finally, a case is shown to illustrate the implementation of the proposed method.","PeriodicalId":251734,"journal":{"name":"2022 3rd International Conference on Computing, Networks and Internet of Things (CNIOT)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122730991","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 : 2022-05-01DOI: 10.1109/cniot55862.2022.00023
B. Xue, Hai Zhao, Wei Yao
Intrusion detection system (IDS) is crucial to security architecture of Internet of Things (IoT). In recent researches, the traditional machine learning and deep learning methods have been applied to the field of intrusion detection and achieved satisfactory performance. However, due to diverse IoT and dynamic network environment, it is difficult to use a single model for heterogeneous IoT networks and collect enough labeled data to train the new model. To solve these issues, we propose an intrusion detection approach based on heterogeneous transfer learning (HTL) for building an intrusion detection model with strong adaptability. Specifically, the approach consists of an Autoencoder architecture for aligning the heterogeneous features and lightweight Convolutional Neural Network (CNN) for unsupervised domain adaptation. Extensive experimental results on three public datasets reveal that the effectiveness of our proposed approach in the IoT environment with unlabeled and limited data.
{"title":"Deep Transfer Learning for IoT Intrusion Detection","authors":"B. Xue, Hai Zhao, Wei Yao","doi":"10.1109/cniot55862.2022.00023","DOIUrl":"https://doi.org/10.1109/cniot55862.2022.00023","url":null,"abstract":"Intrusion detection system (IDS) is crucial to security architecture of Internet of Things (IoT). In recent researches, the traditional machine learning and deep learning methods have been applied to the field of intrusion detection and achieved satisfactory performance. However, due to diverse IoT and dynamic network environment, it is difficult to use a single model for heterogeneous IoT networks and collect enough labeled data to train the new model. To solve these issues, we propose an intrusion detection approach based on heterogeneous transfer learning (HTL) for building an intrusion detection model with strong adaptability. Specifically, the approach consists of an Autoencoder architecture for aligning the heterogeneous features and lightweight Convolutional Neural Network (CNN) for unsupervised domain adaptation. Extensive experimental results on three public datasets reveal that the effectiveness of our proposed approach in the IoT environment with unlabeled and limited data.","PeriodicalId":251734,"journal":{"name":"2022 3rd International Conference on Computing, Networks and Internet of Things (CNIOT)","volume":"106 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127529555","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 : 2022-05-01DOI: 10.1109/cniot55862.2022.00022
J. Li, Heng Liu, Xinhua Lu, Shanwen Guan, Mengge Li
Beacon indoor positioning methods with Bluetooth has attracted the interest of researchers by its low cost, low energy consumption, and easy implementation. The key two steps of positioning in wireless systems are inference the accurate distance from the received signal strength indication (RSSI), and computing accurate location information from the inferred distance. The noise and multi-paths of the wireless channel will lead to complex nonlinearity relationship between RSSI and distance, which is difficult to modeled directly by the simple functions. Back Propagation (BP) neural networks can be used to construct ranging models with RSSI, but it will easily fall into local optimization which will lead to inaccurate. Besides, the MinMax localization algorithm used in the computing location information is easily affected by the fluctuation of the range value then will affect the localization accuracy. In this paper, a combined dynamic adaptive sparrow search (DASSA) with two-step residual network optimized MinMax (TSRK-MinMax) algorithm is proposed to improve the accuracy of beacon indoor localization. First, we use a dynamic adaptive sparrow search (DASSA) algorithm to optimize BP neural network to improve the ranging accuracy. Next, we utilize a K-Nearest Neighbor(KNN) algorithm to select the base stations which can receive the best signal. Then a two-step residual network are used to optimized MinMax algorithm (TSRK-MinMax) which leads to accuracy localization. The experimental results show that the overall localization error is reduced by 11.7%, which effectively improves the accuracy and robustness of the Beacon indoor localization.
{"title":"A combined DASSA-BP and TSRK-MinMax algorithm for high accuracy beacon indoor positioning","authors":"J. Li, Heng Liu, Xinhua Lu, Shanwen Guan, Mengge Li","doi":"10.1109/cniot55862.2022.00022","DOIUrl":"https://doi.org/10.1109/cniot55862.2022.00022","url":null,"abstract":"Beacon indoor positioning methods with Bluetooth has attracted the interest of researchers by its low cost, low energy consumption, and easy implementation. The key two steps of positioning in wireless systems are inference the accurate distance from the received signal strength indication (RSSI), and computing accurate location information from the inferred distance. The noise and multi-paths of the wireless channel will lead to complex nonlinearity relationship between RSSI and distance, which is difficult to modeled directly by the simple functions. Back Propagation (BP) neural networks can be used to construct ranging models with RSSI, but it will easily fall into local optimization which will lead to inaccurate. Besides, the MinMax localization algorithm used in the computing location information is easily affected by the fluctuation of the range value then will affect the localization accuracy. In this paper, a combined dynamic adaptive sparrow search (DASSA) with two-step residual network optimized MinMax (TSRK-MinMax) algorithm is proposed to improve the accuracy of beacon indoor localization. First, we use a dynamic adaptive sparrow search (DASSA) algorithm to optimize BP neural network to improve the ranging accuracy. Next, we utilize a K-Nearest Neighbor(KNN) algorithm to select the base stations which can receive the best signal. Then a two-step residual network are used to optimized MinMax algorithm (TSRK-MinMax) which leads to accuracy localization. The experimental results show that the overall localization error is reduced by 11.7%, which effectively improves the accuracy and robustness of the Beacon indoor localization.","PeriodicalId":251734,"journal":{"name":"2022 3rd International Conference on Computing, Networks and Internet of Things (CNIOT)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133096071","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 : 2022-05-01DOI: 10.1109/cniot55862.2022.00017
Qiyan Zhan, Nan Liu, Zhiwen Pan, Hongjian Sun
This paper studies the observability of the power grid by jointly considering the power system with the wireless communication system under the strict latency requirements of Phasor Measurement Units (PMUs), which is characterized via the theory of effective capacity. In order to meet the quality of service (QoS) requirements and save communication bandwidth at the same time, the technique of non-orthogonal multiple access (NOMA) is adopted. For practical purposes, we consider the case where each NOMA group consists of at most two PMUs. The problem is formulated as minimizing the required communication bandwidth while satisfying the observability constraint of the power grid, over all possible NOMA user pairing strategies, bandwidth allocation among NOMA groups, power allocation within each NOMA group and the normalized QoS exponents of each PMU. In order to solve this problem, we first derive the closed-form expressions of the effective capacity of the two-PMU NOMA pair, then the problem is solved by first fixing the NOMA user pairing strategy and the probability of successful transmission of each PMU, and finding the optimal bandwidth allocation among NOMA groups, power allocation within each NOMA group and the normalized QoS exponents of each PMU via the bisection search method. Then, the probability of successful transmission of each PMU is found via the simulated annealing algorithm for a fixed NOMA user pairing strategy. Finally, we discuss the benefit of using uniform channel gain difference (UCGD) pairing as the PMU-pairing strategy. Numerical results on the IEEE 14-bus power system demonstrate the significant bandwidth savings of the proposed algorithm compared with the orthogonal multiple access (OMA) method and the NOMA method with equal bandwidth and power allocation.
{"title":"Statistical Power Grid Observability under NOMA-based Communication Constraints","authors":"Qiyan Zhan, Nan Liu, Zhiwen Pan, Hongjian Sun","doi":"10.1109/cniot55862.2022.00017","DOIUrl":"https://doi.org/10.1109/cniot55862.2022.00017","url":null,"abstract":"This paper studies the observability of the power grid by jointly considering the power system with the wireless communication system under the strict latency requirements of Phasor Measurement Units (PMUs), which is characterized via the theory of effective capacity. In order to meet the quality of service (QoS) requirements and save communication bandwidth at the same time, the technique of non-orthogonal multiple access (NOMA) is adopted. For practical purposes, we consider the case where each NOMA group consists of at most two PMUs. The problem is formulated as minimizing the required communication bandwidth while satisfying the observability constraint of the power grid, over all possible NOMA user pairing strategies, bandwidth allocation among NOMA groups, power allocation within each NOMA group and the normalized QoS exponents of each PMU. In order to solve this problem, we first derive the closed-form expressions of the effective capacity of the two-PMU NOMA pair, then the problem is solved by first fixing the NOMA user pairing strategy and the probability of successful transmission of each PMU, and finding the optimal bandwidth allocation among NOMA groups, power allocation within each NOMA group and the normalized QoS exponents of each PMU via the bisection search method. Then, the probability of successful transmission of each PMU is found via the simulated annealing algorithm for a fixed NOMA user pairing strategy. Finally, we discuss the benefit of using uniform channel gain difference (UCGD) pairing as the PMU-pairing strategy. Numerical results on the IEEE 14-bus power system demonstrate the significant bandwidth savings of the proposed algorithm compared with the orthogonal multiple access (OMA) method and the NOMA method with equal bandwidth and power allocation.","PeriodicalId":251734,"journal":{"name":"2022 3rd International Conference on Computing, Networks and Internet of Things (CNIOT)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134014673","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 : 2022-05-01DOI: 10.1109/cniot55862.2022.00026
Qiyan Zhan, Nan Liu, Zhiwen Pan, Hongjian Sun
We study the stochastic observability of the power grid system under communication constraints in the finite blocklength regime. Compared to the study under the assumption of infinite blocklength, we introduce two new elements: probability of decoding error and transmission delay. An optimization problem to maximize the observability of the smart grid over all possible bandwidth allocation is proposed, incorporating these two new elements. To solve the optimization problem, for a given bandwidth allocation, we first solve parallel subproblems, one for each synchronous phasor measurement unit (PMU), using alternating optimization, to find the optimal QoS exponent, transmission delay and probability of decoding error for each PMU. Then, simulated annealing method is used to find the optimal bandwidth allocation among PMUs. Numerical results verify that the assumption of infinite blocklength is indeed too optimistic and instead, finite blocklength should be studied. Large bandwidth saving gains of the proposed scheme are demonstrated compared to the equal bandwidth allocation scheme.
{"title":"Statistical Power Grid Observability under Finite Blocklength","authors":"Qiyan Zhan, Nan Liu, Zhiwen Pan, Hongjian Sun","doi":"10.1109/cniot55862.2022.00026","DOIUrl":"https://doi.org/10.1109/cniot55862.2022.00026","url":null,"abstract":"We study the stochastic observability of the power grid system under communication constraints in the finite blocklength regime. Compared to the study under the assumption of infinite blocklength, we introduce two new elements: probability of decoding error and transmission delay. An optimization problem to maximize the observability of the smart grid over all possible bandwidth allocation is proposed, incorporating these two new elements. To solve the optimization problem, for a given bandwidth allocation, we first solve parallel subproblems, one for each synchronous phasor measurement unit (PMU), using alternating optimization, to find the optimal QoS exponent, transmission delay and probability of decoding error for each PMU. Then, simulated annealing method is used to find the optimal bandwidth allocation among PMUs. Numerical results verify that the assumption of infinite blocklength is indeed too optimistic and instead, finite blocklength should be studied. Large bandwidth saving gains of the proposed scheme are demonstrated compared to the equal bandwidth allocation scheme.","PeriodicalId":251734,"journal":{"name":"2022 3rd International Conference on Computing, Networks and Internet of Things (CNIOT)","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114713039","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 : 2022-05-01DOI: 10.1109/cniot55862.2022.00010
Tianjian Yang, Qiao Sun, Feng Chen
Hybrid transaction in blockchain enlarges the application domain of smart contracts by enabling off-chain data access. In support of the hybrid transaction, blockchain oracle is an indispensable middleware responsible for receiving data-request tasks and pushing the result back to chain. Though solutions supporting hybrid transactions do exist, they are less featured or applicable for both contract and application developers. On one hand, the racing-condition among several concurrent hybrid transactions should be manually resolved by contract developers. On the other hand, application developers have to associate each individual task with its parent hybrid transaction in order to obtain the result. In this paper, a novel block-chain oracle platform with a set of simple but unified API for both hybrid and traditional transactions is proposed. Firstly, we design and implement a transactional storage layer and a transaction scheduler for transaction preservation and ordering. Secondly, with the help of an automatic result tracing mechanism, application developers are able to acquire transaction results transparently without awareness of oracles and underlying task execution. Experiments are conducted on one of current prevalent permissioned blockchain platforms, RepChain, and the result shows that all of the system design goals are fulfilled with merely an acceptable loss of performance.
{"title":"TransOra: A Transaction-preserving and Transparent Distributed Oracle on Permissioned Blockchain For Hybrid Smart Contracts","authors":"Tianjian Yang, Qiao Sun, Feng Chen","doi":"10.1109/cniot55862.2022.00010","DOIUrl":"https://doi.org/10.1109/cniot55862.2022.00010","url":null,"abstract":"Hybrid transaction in blockchain enlarges the application domain of smart contracts by enabling off-chain data access. In support of the hybrid transaction, blockchain oracle is an indispensable middleware responsible for receiving data-request tasks and pushing the result back to chain. Though solutions supporting hybrid transactions do exist, they are less featured or applicable for both contract and application developers. On one hand, the racing-condition among several concurrent hybrid transactions should be manually resolved by contract developers. On the other hand, application developers have to associate each individual task with its parent hybrid transaction in order to obtain the result. In this paper, a novel block-chain oracle platform with a set of simple but unified API for both hybrid and traditional transactions is proposed. Firstly, we design and implement a transactional storage layer and a transaction scheduler for transaction preservation and ordering. Secondly, with the help of an automatic result tracing mechanism, application developers are able to acquire transaction results transparently without awareness of oracles and underlying task execution. Experiments are conducted on one of current prevalent permissioned blockchain platforms, RepChain, and the result shows that all of the system design goals are fulfilled with merely an acceptable loss of performance.","PeriodicalId":251734,"journal":{"name":"2022 3rd International Conference on Computing, Networks and Internet of Things (CNIOT)","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131866265","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 : 2022-04-26DOI: 10.48550/arXiv.2204.13589
Wei Zhong Tee, Rushit Dave, Naeem Seliya, Mounika Vanamala
Human activity recognition using deep learning techniques has become increasing popular because of its high effectivity with recognizing complex tasks, as well as being relatively low in costs compared to more traditional machine learning techniques. This paper surveys some state-of-the-art human activity recognition models that are based on deep learning architecture and has layers containing Convolution Neural Networks (CNN), Long Short-Term Memory (LSTM), or a mix of more than one type for a hybrid system. The analysis outlines how the models are implemented to maximize its effectivity and some of the potential limitations it faces. Keywords: Human Activity Recognition, Deep Learning
{"title":"A Close Look into Human Activity Recognition Models using Deep Learning","authors":"Wei Zhong Tee, Rushit Dave, Naeem Seliya, Mounika Vanamala","doi":"10.48550/arXiv.2204.13589","DOIUrl":"https://doi.org/10.48550/arXiv.2204.13589","url":null,"abstract":"Human activity recognition using deep learning techniques has become increasing popular because of its high effectivity with recognizing complex tasks, as well as being relatively low in costs compared to more traditional machine learning techniques. This paper surveys some state-of-the-art human activity recognition models that are based on deep learning architecture and has layers containing Convolution Neural Networks (CNN), Long Short-Term Memory (LSTM), or a mix of more than one type for a hybrid system. The analysis outlines how the models are implemented to maximize its effectivity and some of the potential limitations it faces. Keywords: Human Activity Recognition, Deep Learning","PeriodicalId":251734,"journal":{"name":"2022 3rd International Conference on Computing, Networks and Internet of Things (CNIOT)","volume":"420 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-04-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"113967130","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：