Pub Date : 2024-11-28DOI: 10.1016/j.compeleceng.2024.109893
Santanu Basak, Kakali Chatterjee
The Federated Learning (FL)-based approaches are increasing rapidly for different areas, such as home automation, smart healthcare, smart cars, etc. In FL, multiple users participate collaboratively and distributively to construct a global model without sharing raw data. The FL-based system resolves several issues of central server-based machine learning approaches, such as data availability, maintaining user privacy, etc. Still, some issues exist, such as data poisoning attacks and re-identification attacks. This paper proposes a Data Poisoning Attack Defense (DPAD) Mechanism that detects and defends against the data poisoning attack efficiently and secures the aggregation process for the Federated Learning-based systems. The DPAD verifies each client’s updates using an audit mechanism that decides whether a local update is considered for aggregation. The experimental results show the effectiveness of the attack and the power of the DPAD mechanism compared with the state-of-the-art methods.
{"title":"DPAD: Data Poisoning Attack Defense Mechanism for federated learning-based system","authors":"Santanu Basak, Kakali Chatterjee","doi":"10.1016/j.compeleceng.2024.109893","DOIUrl":"10.1016/j.compeleceng.2024.109893","url":null,"abstract":"<div><div>The Federated Learning (FL)-based approaches are increasing rapidly for different areas, such as home automation, smart healthcare, smart cars, etc. In FL, multiple users participate collaboratively and distributively to construct a global model without sharing raw data. The FL-based system resolves several issues of central server-based machine learning approaches, such as data availability, maintaining user privacy, etc. Still, some issues exist, such as data poisoning attacks and re-identification attacks. This paper proposes a Data Poisoning Attack Defense (DPAD) Mechanism that detects and defends against the data poisoning attack efficiently and secures the aggregation process for the Federated Learning-based systems. The DPAD verifies each client’s updates using an audit mechanism that decides whether a local update is considered for aggregation. The experimental results show the effectiveness of the attack and the power of the DPAD mechanism compared with the state-of-the-art methods.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"121 ","pages":"Article 109893"},"PeriodicalIF":4.0,"publicationDate":"2024-11-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142747894","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-28DOI: 10.1016/j.compeleceng.2024.109916
Zhiwei Song , Xinbo Huang , Chao Ji , Ye Zhang , Zhang Chao , Yang Peng
The identification of steel strip defects plays a pivotal role in assessing steel quality and advancing production technology. However, the majority of intelligent defect recognition algorithms for steel strips, based on deep learning, primarily focus on supervised learning. These methods depend on a multitude of training samples, incurring additional manual labelling costs, and exhibit low recognition efficiency. In contrast to supervised learning, we integrate the fine-grained characteristics of strip defects. We propose a cross-domain, fine-grained strip defect detection method based on semi-supervised learning and Multi-head self-attention coordination, along with an improvement strategy, resulting in a novel network structure: Multi-head Self Attention and Semi-supervised collaborative detection network (MSD Net). This method initiates the cross-domain migration of defect samples through Cycle Generative Adversarial Networks (Cycle GAN), creating new semi-supervised training samples from source domain and target domain data to enhance data distribution diversity. The detection model is then constructed leveraging the advantages of Multi-head Self Attention (MSA) in augmenting the global receptive field of feature extraction. The proposed semi-supervised learning method employs a pseudo-label allocation strategy to guide the model in fully utilizing the distribution fitting of unlabeled samples. This allows the deep neural network to learn a more comprehensive multivariate data distribution within the training set, thereby enhancing the generalization ability of the semi-supervised model. Experimental results on the benchmark dataset for steel strip defect detection demonstrate that the cross-domain semi-supervised method achieves a test accuracy of 96.1 % on mAP@0.5, surpassing the supervised baseline model by 4.8 %. Our method also outperforms the baseline supervised model in the accuracy of small target recognition on PASCAL VOC 2007 datasets. Additionally, we have implemented a strip defect detection system based on edge computing for real-time deployment of the proposed algorithm. Testing in an actual industrial setting further validates the efficacy of our proposed method in practical applications. Our work encourages further exploration, the task of public datasets can be obtained at https://github.com/songzhiweiknight/NEU-DET-Datasets.git.
{"title":"Cross-domain fine grained strip steel defect detection method based on semi-supervised learning and Multi-head Self Attention coordination","authors":"Zhiwei Song , Xinbo Huang , Chao Ji , Ye Zhang , Zhang Chao , Yang Peng","doi":"10.1016/j.compeleceng.2024.109916","DOIUrl":"10.1016/j.compeleceng.2024.109916","url":null,"abstract":"<div><div>The identification of steel strip defects plays a pivotal role in assessing steel quality and advancing production technology. However, the majority of intelligent defect recognition algorithms for steel strips, based on deep learning, primarily focus on supervised learning. These methods depend on a multitude of training samples, incurring additional manual labelling costs, and exhibit low recognition efficiency. In contrast to supervised learning, we integrate the fine-grained characteristics of strip defects. We propose a cross-domain, fine-grained strip defect detection method based on semi-supervised learning and Multi-head self-attention coordination, along with an improvement strategy, resulting in a novel network structure: Multi-head Self Attention and Semi-supervised collaborative detection network (MSD Net). This method initiates the cross-domain migration of defect samples through Cycle Generative Adversarial Networks (Cycle GAN), creating new semi-supervised training samples from source domain and target domain data to enhance data distribution diversity. The detection model is then constructed leveraging the advantages of Multi-head Self Attention (MSA) in augmenting the global receptive field of feature extraction. The proposed semi-supervised learning method employs a pseudo-label allocation strategy to guide the model in fully utilizing the distribution fitting of unlabeled samples. This allows the deep neural network to learn a more comprehensive multivariate data distribution within the training set, thereby enhancing the generalization ability of the semi-supervised model. Experimental results on the benchmark dataset for steel strip defect detection demonstrate that the cross-domain semi-supervised method achieves a test accuracy of 96.1 % on mAP<sup>@0.5</sup>, surpassing the supervised baseline model by 4.8 %. Our method also outperforms the baseline supervised model in the accuracy of small target recognition on PASCAL VOC 2007 datasets. Additionally, we have implemented a strip defect detection system based on edge computing for real-time deployment of the proposed algorithm. Testing in an actual industrial setting further validates the efficacy of our proposed method in practical applications. Our work encourages further exploration, the task of public datasets can be obtained at <span><span>https://github.com/songzhiweiknight/NEU-DET-Datasets.git</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"121 ","pages":"Article 109916"},"PeriodicalIF":4.0,"publicationDate":"2024-11-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142747992","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-27DOI: 10.1016/j.compeleceng.2024.109892
Rafael Ferreira, Ivo Bispo, Carlos Rabadão, Leonel Santos, Rogério Luís de C. Costa
In recent years, the Internet of Things (IoT) revolutionized agricultural management by enabling data-driven decision-making through seamless connectivity among various devices and equipment. The security of Agricultural IoT (AG-IoT) devices becomes increasingly evident as reliance on them grows. On the other hand, machine learning models for intrusion detection show promise in identifying vulnerabilities, but their effectiveness depends on being trained on representative data. Indeed, there is a notable gap in network intrusion detection for AG-IoT, as existing datasets for training machine learning models lack the context of AG-IoT scenarios. Also, most existing ones rely on packed-based features (and not on network flow data), and analysing such data can be resource-consuming.
In this work, we present the “Farm-Flow” dataset. We created a realistic AG-IoT scenario to build the dataset and executed eight types of network attacks. Over one million instances of relevant data were collected, which we combined into network flows, organized and made publicly available via http://doi.org/10.5281/zenodo.10964647.
The dataset created has been evaluated using multiple intrusion detection models in terms of their capabilities to identify and classify malicious traffic. The assessed models presented high performance and even achieved an F1-score of more than 90% while identifying malicious traffic. The “Farm-Flow” may support the training of intrusion detection methods, and the performance results contribute to future benchmarking.
{"title":"Farm-flow dataset: Intrusion detection in smart agriculture based on network flows","authors":"Rafael Ferreira, Ivo Bispo, Carlos Rabadão, Leonel Santos, Rogério Luís de C. Costa","doi":"10.1016/j.compeleceng.2024.109892","DOIUrl":"10.1016/j.compeleceng.2024.109892","url":null,"abstract":"<div><div>In recent years, the Internet of Things (IoT) revolutionized agricultural management by enabling data-driven decision-making through seamless connectivity among various devices and equipment. The security of Agricultural IoT (AG-IoT) devices becomes increasingly evident as reliance on them grows. On the other hand, machine learning models for intrusion detection show promise in identifying vulnerabilities, but their effectiveness depends on being trained on representative data. Indeed, there is a notable gap in network intrusion detection for AG-IoT, as existing datasets for training machine learning models lack the context of AG-IoT scenarios. Also, most existing ones rely on packed-based features (and not on network flow data), and analysing such data can be resource-consuming.</div><div>In this work, we present the “Farm-Flow” dataset. We created a realistic AG-IoT scenario to build the dataset and executed eight types of network attacks. Over one million instances of relevant data were collected, which we combined into network flows, organized and made publicly available via <span><span>http://doi.org/10.5281/zenodo.10964647</span><svg><path></path></svg></span>.</div><div>The dataset created has been evaluated using multiple intrusion detection models in terms of their capabilities to identify and classify malicious traffic. The assessed models presented high performance and even achieved an F1-score of more than 90% while identifying malicious traffic. The “Farm-Flow” may support the training of intrusion detection methods, and the performance results contribute to future benchmarking.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"121 ","pages":"Article 109892"},"PeriodicalIF":4.0,"publicationDate":"2024-11-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142720469","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Modern mobile applications harvest many user-generated events during execution using proper libraries called analytic libraries. The collection of such events allows the app developers to acquire helpful information to further improve the app. The same collected events are likewise an essential source of information for analytic library providers (e.g., Google and Meta) to understand users’ preferences. However, the user is not involved in this process. To counteract this problem, some proposals arose from legal (e.g., General Data Protection Regulation (GDPR)) and research perspectives. Concerning the latter point, some research efforts led to the definition of solutions for the Android ecosystem that allow one to limit the gathering of such data before the analytic libraries collect it or give the user control of the process. To this aim, HideDroid was the first proposal to allow the user to define different privacy levels for each app installed on the device by leveraging k-anonymity and differential privacy techniques. Subsequently, VirtualHideDroid extended HideDroid by taking advantage of the same approach to virtualized Android environments, in which an application (plugin) can run within another application (container). In this scenario, VirtualHideDroid anonymizes user event data running as the container app. However, according to standard threat models regarding virtualized Android environments, assuming that the container app is fully trusted is too optimistic in real deployments.
For this reason, in this paper, we extend the work of the original VirtualHideDroid work by assuming that the same tool may be untrusted, i.e., controlled by an external attacker that has access to the container app, thereby having full access to the user data. To solve this problem, we define a new approach, named MarvelHideDroid, which gives reliable anonymization of event data in the Plugin app, even in the event of a malicious/compromised container. Moreover, and differently from VirtualHideDroid, MarvelHideDroid relies on LLM to automatically build up the generalizations required by k-anonymity, resulting in an anonymization strategy that is more reliable against modification in the data structure of the events captured by the analytic libraries. We empirically demonstrate the viability and reliability of the proposal by testing an implementation of MarvelHideDroid on a set of real Android apps in a virtualized environment.
{"title":"MarvelHideDroid: Reliable on-the-fly data anonymization based on Android virtualization","authors":"Francesco Pagano , Luca Verderame , Enrico Russo , Alessio Merlo","doi":"10.1016/j.compeleceng.2024.109882","DOIUrl":"10.1016/j.compeleceng.2024.109882","url":null,"abstract":"<div><div>Modern mobile applications harvest many user-generated events during execution using proper libraries called <em>analytic libraries</em>. The collection of such events allows the app developers to acquire helpful information to further improve the app. The same collected events are likewise an essential source of information for analytic library providers (e.g., Google and Meta) to understand users’ preferences. However, the user is not involved in this process. To counteract this problem, some proposals arose from legal (e.g., General Data Protection Regulation (GDPR)) and research perspectives. Concerning the latter point, some research efforts led to the definition of solutions for the Android ecosystem that allow one to limit the gathering of such data before the analytic libraries collect it or give the user control of the process. To this aim, <em>HideDroid</em> was the first proposal to allow the user to define different privacy levels for each app installed on the device by leveraging k-anonymity and differential privacy techniques. Subsequently, <em>VirtualHideDroid</em> extended HideDroid by taking advantage of the same approach to virtualized Android environments, in which an application (plugin) can run within another application (container). In this scenario, VirtualHideDroid anonymizes user event data running as the container app. However, according to standard threat models regarding virtualized Android environments, assuming that the container app is fully trusted is too optimistic in real deployments.</div><div>For this reason, in this paper, we extend the work of the original VirtualHideDroid work by assuming that the same tool may be untrusted, i.e., controlled by an external attacker that has access to the container app, thereby having full access to the user data. To solve this problem, we define a new approach, named <em>MarvelHideDroid</em>, which gives reliable anonymization of event data in the Plugin app, even in the event of a malicious/compromised container. Moreover, and differently from VirtualHideDroid, <em>MarvelHideDroid</em> relies on LLM to automatically build up the generalizations required by k-anonymity, resulting in an anonymization strategy that is more reliable against modification in the data structure of the events captured by the analytic libraries. We empirically demonstrate the viability and reliability of the proposal by testing an implementation of <em>MarvelHideDroid</em> on a set of real Android apps in a virtualized environment.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"121 ","pages":"Article 109882"},"PeriodicalIF":4.0,"publicationDate":"2024-11-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142720541","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-27DOI: 10.1016/j.compeleceng.2024.109856
Zuhua Dai, Yifu He
Intelligent classification of software requirements is a hot research issue in the field of requirements engineering. Complete and accurate identification of functional requirements (FRs) and non-functional requirements (NFRs) is the primary task of requirements engineering. However, in real software projects, NFRs are easily neglected and may become a potential risk of project failure. Text is the main source of information about software requirements. With the increasing scale of software projects, a large number of complex types of text materials are used for software requirements analysis. Manual identification of NFRs of software projects has the problems of easy omission, ambiguity, vagueness, and high time-consuming cost. Based on the above existing defects, a deep neural network model named CAFLE is designed in this paper to solve it. CAFLE is composed of two parts, Text-label cooperative attention encoder (TLCAE) and Label decoder (LD). TLCAE adopts a Bi-directional long short-term memory network (Bi-LSTM) and multi-head cooperative attention mechanism to generate an encoded representation of the mutual involvement of requirement classification labels and requirement text. LD is an LSTM decoder with an attention mechanism constructed for the multi-class classification task of requirement text. LD utilizes the representation generated by TLCAE for prediction. Experimental results on the PROMISE benchmark dataset show that CAFLE outperforms existing NFRs classification methods with an F1 score of 95%.
{"title":"A non-functional requirements classification model based on cooperative attention mechanism fused with label embedding","authors":"Zuhua Dai, Yifu He","doi":"10.1016/j.compeleceng.2024.109856","DOIUrl":"10.1016/j.compeleceng.2024.109856","url":null,"abstract":"<div><div>Intelligent classification of software requirements is a hot research issue in the field of requirements engineering. Complete and accurate identification of functional requirements (FRs) and non-functional requirements (NFRs) is the primary task of requirements engineering. However, in real software projects, NFRs are easily neglected and may become a potential risk of project failure. Text is the main source of information about software requirements. With the increasing scale of software projects, a large number of complex types of text materials are used for software requirements analysis. Manual identification of NFRs of software projects has the problems of easy omission, ambiguity, vagueness, and high time-consuming cost. Based on the above existing defects, a deep neural network model named CAFLE is designed in this paper to solve it. CAFLE is composed of two parts, Text-label cooperative attention encoder (TLCAE) and Label decoder (LD). TLCAE adopts a Bi-directional long short-term memory network (Bi-LSTM) and multi-head cooperative attention mechanism to generate an encoded representation of the mutual involvement of requirement classification labels and requirement text. LD is an LSTM decoder with an attention mechanism constructed for the multi-class classification task of requirement text. LD utilizes the representation generated by TLCAE for prediction. Experimental results on the PROMISE benchmark dataset show that CAFLE outperforms existing NFRs classification methods with an F1 score of 95%.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"121 ","pages":"Article 109856"},"PeriodicalIF":4.0,"publicationDate":"2024-11-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142720539","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-27DOI: 10.1016/j.compeleceng.2024.109883
Sakorn Mekruksavanich , Anuchit Jitpattanakul
User identification is a critical aspect of smartwatch security, ensuring that only authorized individuals gain access to sensitive information stored on the device. Conventional methods like passwords and biometrics have limitations, such as the risk of forgetting passwords or the potential for biometric data to be compromised. This research proposes a novel approach for user identification on smartwatches by analyzing activity patterns using a hybrid residual neural network called Att-ResBiLSTM. The proposed method leverages unique patterns of user interactions with their smartwatches, including application usage, typing behavior, and motion sensor data, to create an individualized user profile. Employing a deep learning network specifically designed for wearable devices, the system can reliably and promptly identify users by analyzing their activity patterns. The Att-ResBiLSTM architecture comprises three key components: convolutional layers, ResBiLSTM, and an attention layer. The convolutional layers extract spatial features from the pre-processed data. At the same time, the ResBiLSTM component captures long-term dependencies in the time-series data by combining the advantages of bidirectional long short-term memory (BiLSTM) and residual connections. The attention mechanism enhances the final recognition features by selectively prioritizing the most informative elements of the input data. The Att-ResBiLSTM model is trained and evaluated using a diverse dataset of user activity patterns. Experimental results demonstrate that the proposed approach achieves remarkable accuracy in user identification, with an accuracy rate of 98.29% and the highest F1-score of 98.24%. The research also conducts a comparative analysis to assess the efficacy of accelerometer data versus gyroscope data, revealing that combining both sensor modalities improves user identification performance. The proposed methodology provides a reliable and user-friendly alternative to conventional user authentication techniques for smartwatches. This approach leverages activity patterns and a hybrid residual deep learning network to offer a robust and efficient solution for user identification based on smartwatch data, thereby enhancing the overall security of wearable devices.
{"title":"A residual deep learning network for smartwatch-based user identification using activity patterns in daily living","authors":"Sakorn Mekruksavanich , Anuchit Jitpattanakul","doi":"10.1016/j.compeleceng.2024.109883","DOIUrl":"10.1016/j.compeleceng.2024.109883","url":null,"abstract":"<div><div>User identification is a critical aspect of smartwatch security, ensuring that only authorized individuals gain access to sensitive information stored on the device. Conventional methods like passwords and biometrics have limitations, such as the risk of forgetting passwords or the potential for biometric data to be compromised. This research proposes a novel approach for user identification on smartwatches by analyzing activity patterns using a hybrid residual neural network called Att-ResBiLSTM. The proposed method leverages unique patterns of user interactions with their smartwatches, including application usage, typing behavior, and motion sensor data, to create an individualized user profile. Employing a deep learning network specifically designed for wearable devices, the system can reliably and promptly identify users by analyzing their activity patterns. The Att-ResBiLSTM architecture comprises three key components: convolutional layers, ResBiLSTM, and an attention layer. The convolutional layers extract spatial features from the pre-processed data. At the same time, the ResBiLSTM component captures long-term dependencies in the time-series data by combining the advantages of bidirectional long short-term memory (BiLSTM) and residual connections. The attention mechanism enhances the final recognition features by selectively prioritizing the most informative elements of the input data. The Att-ResBiLSTM model is trained and evaluated using a diverse dataset of user activity patterns. Experimental results demonstrate that the proposed approach achieves remarkable accuracy in user identification, with an accuracy rate of 98.29% and the highest F1-score of 98.24%. The research also conducts a comparative analysis to assess the efficacy of accelerometer data versus gyroscope data, revealing that combining both sensor modalities improves user identification performance. The proposed methodology provides a reliable and user-friendly alternative to conventional user authentication techniques for smartwatches. This approach leverages activity patterns and a hybrid residual deep learning network to offer a robust and efficient solution for user identification based on smartwatch data, thereby enhancing the overall security of wearable devices.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"121 ","pages":"Article 109883"},"PeriodicalIF":4.0,"publicationDate":"2024-11-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142720540","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-26DOI: 10.1016/j.compeleceng.2024.109924
Zhan Li, Xingyu Song, Shi Chen, Kazuyuki Demachi
Nowadays, the technologies in computer vision (CV) are labor-saving and convenient to identify human malicious behaviors. However, they usually fail to consider the robustness, generalization and interpretability of calculation frameworks. In this paper, a very common but sometimes difficult-to-detect case research called armed boundary sabotage is conducted, which is achieved by computer vision module (CVM) and reasoning module (RM). Among them, CVM is used for extracting the key information from raw videos, while RM is applied to obtain the final reasoning results. Considering the transient and confusing properties in such scenarios, a specific human-object interaction analysis process with soft constraint is proposed in CVM. In addition, two reasoning methods which are data-based reasoning method and language-based reasoning methods are implemented in RM. The results show that the human-object interaction analysis process with soft constraint prove to be effective and practical, while the optimal testing accuracy achieves 0.7871. Furthermore, the two proposed reasoning methods are promising for identification of human malicious behaviors. Among them, the advanced language-based reasoning method outperforms others, with highest precision value of 0.8750 and perfect recall value of 1.0000. Besides, these proposals are also verified to be high-performance in other external intrusion scenarios of our previous work. Finally, our research also obtain state-of-the-art results by comparing with other related works.
{"title":"Armed boundary sabotage: A case study of human malicious behaviors identification with computer vision and explainable reasoning methods","authors":"Zhan Li, Xingyu Song, Shi Chen, Kazuyuki Demachi","doi":"10.1016/j.compeleceng.2024.109924","DOIUrl":"10.1016/j.compeleceng.2024.109924","url":null,"abstract":"<div><div>Nowadays, the technologies in computer vision (CV) are labor-saving and convenient to identify human malicious behaviors. However, they usually fail to consider the robustness, generalization and interpretability of calculation frameworks. In this paper, a very common but sometimes difficult-to-detect case research called armed boundary sabotage is conducted, which is achieved by computer vision module (CVM) and reasoning module (RM). Among them, CVM is used for extracting the key information from raw videos, while RM is applied to obtain the final reasoning results. Considering the transient and confusing properties in such scenarios, a specific human-object interaction analysis process with soft constraint is proposed in CVM. In addition, two reasoning methods which are data-based reasoning method and language-based reasoning methods are implemented in RM. The results show that the human-object interaction analysis process with soft constraint prove to be effective and practical, while the optimal testing accuracy achieves 0.7871. Furthermore, the two proposed reasoning methods are promising for identification of human malicious behaviors. Among them, the advanced language-based reasoning method outperforms others, with highest precision value of 0.8750 and perfect recall value of 1.0000. Besides, these proposals are also verified to be high-performance in other external intrusion scenarios of our previous work. Finally, our research also obtain state-of-the-art results by comparing with other related works.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"121 ","pages":"Article 109924"},"PeriodicalIF":4.0,"publicationDate":"2024-11-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142720538","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Speaker identification is a method of identifying an individual from a set of speakers, and text-independent speaker identification systems allow speakers to utter any phrase without any constraints. This study is focused on raw audio analysis as phase, fine-grained frequency patterns, timing cues, and other minute characteristics are preserved when raw waveforms are processed as compared to handcrafted features like Mel-Frequency Cepstral Coefficients (MFCC) and visual representation of audio-like spectrogram. Due to the depth of information, which includes variations in speech rhythm, pitch, and vocal tract shape, it is beneficial for identifying speakers. The deep learning architecture known as SincNet has gained popularity in speaker identification because of its parametric Sinc functions that allow it to operate directly on the raw audio input. In this paper, we have considered SincNet as the baseline model for speaker identification. The effect of proper speech boundary detection, including high-level features and effective optimizer selection are analysed. The precise identification of the signal start and terminus point is important for eliminating the redundant non-speech regions. We have included endpoint detection module as a pre-processing step in the system. Proper feature extraction and selection are crucial to the model's success. To extract more abstract features from the data, we have added more convolution layers to the original SincNet model. Further, we investigated the hyperparameter tuning protocol's sensitivity to the optimizer and selected the suitable optimizer for raw audio analysis. With all the modifications in the system architecture, we are able to archive improvements of 12.76 %, 13.33 %, and 13.39 % respectively for training, validation, and testing over the original SincNet model. In terms of validation loss, our proposed approach attains 0.35 in comparison to the original SincNet loss of 1.02. With this significant improvement, the total training time is marginally increased by 20 minutes for our proposed model. We have performed our investigation on the LibriSpeech dataset to check the effectiveness of our proposed system in comparison to the other model..
{"title":"Text-independent speaker identification using modified SincNet with robust features from suitable acoustic region and appropriate optimizer for raw audio analysis","authors":"Nirupam Shome , Richik Kashyap , Rabul Hussain Laskar","doi":"10.1016/j.compeleceng.2024.109915","DOIUrl":"10.1016/j.compeleceng.2024.109915","url":null,"abstract":"<div><div>Speaker identification is a method of identifying an individual from a set of speakers, and text-independent speaker identification systems allow speakers to utter any phrase without any constraints. This study is focused on raw audio analysis as phase, fine-grained frequency patterns, timing cues, and other minute characteristics are preserved when raw waveforms are processed as compared to handcrafted features like Mel-Frequency Cepstral Coefficients (MFCC) and visual representation of audio-like spectrogram. Due to the depth of information, which includes variations in speech rhythm, pitch, and vocal tract shape, it is beneficial for identifying speakers. The deep learning architecture known as SincNet has gained popularity in speaker identification because of its parametric Sinc functions that allow it to operate directly on the raw audio input. In this paper, we have considered SincNet as the baseline model for speaker identification. The effect of proper speech boundary detection, including high-level features and effective optimizer selection are analysed. The precise identification of the signal start and terminus point is important for eliminating the redundant non-speech regions. We have included endpoint detection module as a pre-processing step in the system. Proper feature extraction and selection are crucial to the model's success. To extract more abstract features from the data, we have added more convolution layers to the original SincNet model. Further, we investigated the hyperparameter tuning protocol's sensitivity to the optimizer and selected the suitable optimizer for raw audio analysis. With all the modifications in the system architecture, we are able to archive improvements of 12.76 %, 13.33 %, and 13.39 % respectively for training, validation, and testing over the original SincNet model. In terms of validation loss, our proposed approach attains 0.35 in comparison to the original SincNet loss of 1.02. With this significant improvement, the total training time is marginally increased by 20 minutes for our proposed model. We have performed our investigation on the LibriSpeech dataset to check the effectiveness of our proposed system in comparison to the other model..</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"121 ","pages":"Article 109915"},"PeriodicalIF":4.0,"publicationDate":"2024-11-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142720491","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
<div><div>In the energy management of a network, it is expected that by extracting the optimal performance for the power sources, storage equipment, and responsive demand, a favorable economic and technology situation is achievable for the network and the mentioned elements. Virtual power plants, as a unit aggregating resources, storage, and responsive loads, can create more favorable conditions in network energy management. So, it is expected that the positive effect of the virtual power plant format on the economic and technical situation of the distribution system is far more than those of managing individual elements mentioned in the network. Consequently, the distribution network operator's economic, environmental, and technical goals are met through the concurrent administration of reactive and active power in the smart distribution network that is equipped with a flexible-sustainable virtual power plant. The system operator is accountable for reducing the weighted sum of the voltage security index, energy loss, and energy cost of the distribution network. This problem is associated with the optimal power flow formulation, which considers the environmental limits and security of voltage in the distribution network, the renewable resource operation model and flexibility in the form of a virtual power plant, and the system's flexibility constraints. Flexibility resources considered in the present study are price-based demand response and electric vehicle parking lots. Stochastic optimization relying on the Unscented Transform assists in providing a suitable model for uncertain quantities resulting from the amount of load, electric vehicles, renewable power, and price of energy and eventually shortens the computing time and accurately computes the flexibility index. The optimal compromise solution amongst various objective functions can be found through fuzzy decision-making. Some innovations of this research include concurrent administration of active and reactive power in virtual power plant, concurrent modeling of economic, operational, environmental, voltage security, and flexibility indicators in the distribution network, utilization of electric vehicles, and demand response as a source of flexibility, use of Unscented transform for modeling the uncertainties corresponding to the exact calculation of flexibility. The suggested method was simulated in the IEEE 69-bus radial smart distribution system. Regarding the numerical report obtained, the optimal performance of each of the renewable generation, demand response, and parking of electric vehicles can significantly impact the economic and technical condition of the distribution network. However, the best condition was obtained when the mentioned elements were placed in the form of a virtual power plant. So, in such a situation, the energy cost is around $1862 for the said network. The lowest value for the worst security index in this network is around 0.933 p.u. Energy loss, maximum volt
{"title":"Eco-power management system with operation and voltage security objectives of distribution system operator considering networked virtual power plants with electric vehicles parking lot and price-based demand response","authors":"Jingyi Zhang , Haotian Wu , Ehsan Akbari , Leila Bagherzadeh , Sasan Pirouzi","doi":"10.1016/j.compeleceng.2024.109895","DOIUrl":"10.1016/j.compeleceng.2024.109895","url":null,"abstract":"<div><div>In the energy management of a network, it is expected that by extracting the optimal performance for the power sources, storage equipment, and responsive demand, a favorable economic and technology situation is achievable for the network and the mentioned elements. Virtual power plants, as a unit aggregating resources, storage, and responsive loads, can create more favorable conditions in network energy management. So, it is expected that the positive effect of the virtual power plant format on the economic and technical situation of the distribution system is far more than those of managing individual elements mentioned in the network. Consequently, the distribution network operator's economic, environmental, and technical goals are met through the concurrent administration of reactive and active power in the smart distribution network that is equipped with a flexible-sustainable virtual power plant. The system operator is accountable for reducing the weighted sum of the voltage security index, energy loss, and energy cost of the distribution network. This problem is associated with the optimal power flow formulation, which considers the environmental limits and security of voltage in the distribution network, the renewable resource operation model and flexibility in the form of a virtual power plant, and the system's flexibility constraints. Flexibility resources considered in the present study are price-based demand response and electric vehicle parking lots. Stochastic optimization relying on the Unscented Transform assists in providing a suitable model for uncertain quantities resulting from the amount of load, electric vehicles, renewable power, and price of energy and eventually shortens the computing time and accurately computes the flexibility index. The optimal compromise solution amongst various objective functions can be found through fuzzy decision-making. Some innovations of this research include concurrent administration of active and reactive power in virtual power plant, concurrent modeling of economic, operational, environmental, voltage security, and flexibility indicators in the distribution network, utilization of electric vehicles, and demand response as a source of flexibility, use of Unscented transform for modeling the uncertainties corresponding to the exact calculation of flexibility. The suggested method was simulated in the IEEE 69-bus radial smart distribution system. Regarding the numerical report obtained, the optimal performance of each of the renewable generation, demand response, and parking of electric vehicles can significantly impact the economic and technical condition of the distribution network. However, the best condition was obtained when the mentioned elements were placed in the form of a virtual power plant. So, in such a situation, the energy cost is around $1862 for the said network. The lowest value for the worst security index in this network is around 0.933 p.u. Energy loss, maximum volt","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"121 ","pages":"Article 109895"},"PeriodicalIF":4.0,"publicationDate":"2024-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142701475","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-25DOI: 10.1016/j.compeleceng.2024.109907
Milad Golzarzadeh , Hashem Oraee , Babak Ganji
Because of exclusive characteristics of Switched Reluctance Motor (SRM) particularly simple and robust structure and high reliability, it can be a good choice for many industrial applications. In terms of structure and performance principles, the Linear Switched Reluctance Motor (LSRM) is similar to rotary SRM (RSRM) and therefore their advantages are identical. The Segmental Translator Linear Switched Reluctance Motor (STLSRM) is a special type of LSRMs that can produce a higher thrust density in comparison to the simple type of LSRM. One of the most important models for predicting the characteristics of electric machines is the model based on the Magnetic Equivalent Circuit (MEC), which can be used to predict the performance characteristics of machines. Although the STLSRM has significant advantages, no MEC model has been introduced for it so far. With the aim of predicting the static and dynamic characteristics of this motor, a new analytical model based on MEC method is developed in the present paper. In addition to simplicity, the developed model has acceptable accuracy and speed. The proposed model is utilized for a three-phase STLSRM and different static and dynamic characteristics of the motor including static flux-linkage, static force, instantaneous current waveform and instantaneous thrust waveform are predicted. To validate these obtained simulation results, they are compared with those derived from the Finite Element Method (FEM).
{"title":"A magnetic equivalent circuit model for segmental translator linear switched reluctance motor","authors":"Milad Golzarzadeh , Hashem Oraee , Babak Ganji","doi":"10.1016/j.compeleceng.2024.109907","DOIUrl":"10.1016/j.compeleceng.2024.109907","url":null,"abstract":"<div><div>Because of exclusive characteristics of Switched Reluctance Motor (SRM) particularly simple and robust structure and high reliability, it can be a good choice for many industrial applications. In terms of structure and performance principles, the Linear Switched Reluctance Motor (LSRM) is similar to rotary SRM (RSRM) and therefore their advantages are identical. The Segmental Translator Linear Switched Reluctance Motor (STLSRM) is a special type of LSRMs that can produce a higher thrust density in comparison to the simple type of LSRM. One of the most important models for predicting the characteristics of electric machines is the model based on the Magnetic Equivalent Circuit (MEC), which can be used to predict the performance characteristics of machines. Although the STLSRM has significant advantages, no MEC model has been introduced for it so far. With the aim of predicting the static and dynamic characteristics of this motor, a new analytical model based on MEC method is developed in the present paper. In addition to simplicity, the developed model has acceptable accuracy and speed. The proposed model is utilized for a three-phase STLSRM and different static and dynamic characteristics of the motor including static flux-linkage, static force, instantaneous current waveform and instantaneous thrust waveform are predicted. To validate these obtained simulation results, they are compared with those derived from the Finite Element Method (FEM).</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"121 ","pages":"Article 109907"},"PeriodicalIF":4.0,"publicationDate":"2024-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142701474","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号