Pub Date : 2024-10-29DOI: 10.1016/j.engappai.2024.109495
This paper introduces a novel architecture for hurricane monitoring aimed at maximizing the collection of critical data to enhance the accuracy of weather predictions. The proposed system deploys a swarm of controllable balloons equipped with meteorological sensors within the hurricane environment. A key challenge in this setup is managing the trade-off between maximizing area coverage for data collection and maintaining robust communication links among the balloons. To address this challenge, we propose a cost function with two conflicting components: one prioritizes area coverage, and the other focuses on repositioning to maintain communication. This cost function is optimized using an adaptive neural network-based model predictive control strategy, which enables the system to dynamically balance these competing requirements in real-time. Quantitative results from extensive simulations demonstrate the versatility and effectiveness of the proposed architecture, showing that it can achieve comprehensive communication connectivity and increased area coverage across various configurations, including different numbers of balloons and operational periods.
{"title":"A novel policy for coordinating a hurricane monitoring system using a swarm of buoyancy-controlled balloons trading off communication and coverage","authors":"","doi":"10.1016/j.engappai.2024.109495","DOIUrl":"10.1016/j.engappai.2024.109495","url":null,"abstract":"<div><div>This paper introduces a novel architecture for hurricane monitoring aimed at maximizing the collection of critical data to enhance the accuracy of weather predictions. The proposed system deploys a swarm of controllable balloons equipped with meteorological sensors within the hurricane environment. A key challenge in this setup is managing the trade-off between maximizing area coverage for data collection and maintaining robust communication links among the balloons. To address this challenge, we propose a cost function with two conflicting components: one prioritizes area coverage, and the other focuses on repositioning to maintain communication. This cost function is optimized using an adaptive neural network-based model predictive control strategy, which enables the system to dynamically balance these competing requirements in real-time. Quantitative results from extensive simulations demonstrate the versatility and effectiveness of the proposed architecture, showing that it can achieve comprehensive communication connectivity and increased area coverage across various configurations, including different numbers of balloons and operational periods.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":null,"pages":null},"PeriodicalIF":7.5,"publicationDate":"2024-10-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142539821","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2024-10-29DOI: 10.1016/j.engappai.2024.109548
As indispensable components of rolling axle boxes, the condition of the bearings affects the safety of the traveling train. Therefore, bearing fault diagnosis is an imperative prerequisite for train safety. However, the diagnosis performance under variable working conditions is degraded owing to the large difference in the sample distribution and fewer samples. Although unsupervised domain adaptation models can solve these problems, environmental noise causes the fault features extracted from the two domains to overlap. Ultimately, the discriminative properties of the different samples remain insufficient. Therefore, we propose a rolling fault diagnosis approach for variable working conditions via lightweight Transformer and homogeneous generalized contrastive learning with inter-class repulsive discriminant (HGCL-ICRD). First, a deformable Transformer with lightweight manner is constructed to extract fault features from historical working conditions. Then, the source domain clustering cluster points are used to construct the positive and negative samples of the target domain to achieve the redistribution of the number. On this basis, the homogeneous generalized contrastive learning approach is built to make the samples to be tested have better classifiability. Finally, an inter-class repulsive discriminant term is constructed to minimize the sample distributional difference between the two domains. Furthermore, we construct an improved gray wolf algorithm to optimize the HGCL-ICRD. Extensive experiments on three datasets demonstrate that our model can perform high-precision and high-efficiency diagnosis under variable working conditions.
{"title":"Bearing fault diagnosis for variable working conditions via lightweight transformer and homogeneous generalized contrastive learning with inter-class repulsive discriminant","authors":"","doi":"10.1016/j.engappai.2024.109548","DOIUrl":"10.1016/j.engappai.2024.109548","url":null,"abstract":"<div><div>As indispensable components of rolling axle boxes, the condition of the bearings affects the safety of the traveling train. Therefore, bearing fault diagnosis is an imperative prerequisite for train safety. However, the diagnosis performance under variable working conditions is degraded owing to the large difference in the sample distribution and fewer samples. Although unsupervised domain adaptation models can solve these problems, environmental noise causes the fault features extracted from the two domains to overlap. Ultimately, the discriminative properties of the different samples remain insufficient. Therefore, we propose a rolling fault diagnosis approach for variable working conditions via lightweight Transformer and homogeneous generalized contrastive learning with inter-class repulsive discriminant (HGCL-ICRD). First, a deformable Transformer with lightweight manner is constructed to extract fault features from historical working conditions. Then, the source domain clustering cluster points are used to construct the positive and negative samples of the target domain to achieve the redistribution of the number. On this basis, the homogeneous generalized contrastive learning approach is built to make the samples to be tested have better classifiability. Finally, an inter-class repulsive discriminant term is constructed to minimize the sample distributional difference between the two domains. Furthermore, we construct an improved gray wolf algorithm to optimize the HGCL-ICRD. Extensive experiments on three datasets demonstrate that our model can perform high-precision and high-efficiency diagnosis under variable working conditions.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":null,"pages":null},"PeriodicalIF":7.5,"publicationDate":"2024-10-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142539820","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2024-10-29DOI: 10.1016/j.engappai.2024.109511
The selection of input variables and their discrete time delays are fundamentally important in developing robust data-driven dynamic models for use in applied engineering settings, such as for controller design or system level calibration/optimisation. This work is not trivial, especially in the case of complex multivariate and non-linear dynamic systems. There are an array of model-free approaches to input selection explored in the literature, including Multivariate Mutual Information (MMI), Gamma Tests (GT), Self-Organising Maps (SOM) and Partial Mutual Information (PMI). Such a filter-based approach has advantages in exploring feature correlations and their associated information content in a data set directly, agnostic to the constraint of any specific model structure or form.
This paper investigates and expands upon the application of a PMI-based Input Selection (PMI-IS) methodology for resulting in a modified version of the algorithm. The modifications are: (1) Selection of input Dead Time (DT) using Mutual Information of First and Second Difference terms of input delays with the output. (2) The Number of Delayed Outputs (NDO) is selected based on the PMI incorporating the previously selected time delays; (3) The Number of Delayed Inputs (NDI) is selected based on the PMI incorporating the identified delay times and NDO; (4) The established Dual-residual PMI (DPMI) algorithm for input selection is simplified to a Single-residual PMI (SPMI) algorithm.
Three benchmark discrete-time non-linear dynamic systems and one practical demonstration are used in the case study to demonstrate the effectiveness of this learning algorithm for data-driven identification of time delays, in addition to the implementation details of this modified SPMI-IS methodology.
{"title":"A Single-residual Partial Mutual Information (SPMI) approach to learning discrete-time inputs of stable nonlinear dynamic systems","authors":"","doi":"10.1016/j.engappai.2024.109511","DOIUrl":"10.1016/j.engappai.2024.109511","url":null,"abstract":"<div><div>The selection of input variables and their discrete time delays are fundamentally important in developing robust data-driven dynamic models for use in applied engineering settings, such as for controller design or system level calibration/optimisation. This work is not trivial, especially in the case of complex multivariate and non-linear dynamic systems. There are an array of model-free approaches to input selection explored in the literature, including Multivariate Mutual Information (MMI), Gamma Tests (GT), Self-Organising Maps (SOM) and Partial Mutual Information (PMI). Such a filter-based approach has advantages in exploring feature correlations and their associated information content in a data set directly, agnostic to the constraint of any specific model structure or form.</div><div>This paper investigates and expands upon the application of a PMI-based Input Selection (PMI-IS) methodology for resulting in a modified version of the algorithm. The modifications are: (1) Selection of input Dead Time (DT) using Mutual Information of First and Second Difference terms of input delays with the output. (2) The Number of Delayed Outputs (NDO) is selected based on the PMI incorporating the previously selected time delays; (3) The Number of Delayed Inputs (NDI) is selected based on the PMI incorporating the identified delay times and NDO; (4) The established Dual-residual PMI (DPMI) algorithm for input selection is simplified to a Single-residual PMI (SPMI) algorithm.</div><div>Three benchmark discrete-time non-linear dynamic systems and one practical demonstration are used in the case study to demonstrate the effectiveness of this learning algorithm for data-driven identification of time delays, in addition to the implementation details of this modified SPMI-IS methodology.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":null,"pages":null},"PeriodicalIF":7.5,"publicationDate":"2024-10-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142539822","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2024-10-28DOI: 10.1016/j.engappai.2024.109518
The utilization of edge devices in fog computing services is increasing every day to achieve effective communication between edge devices as it reduces the latency and processing time. When the number of edge devices increases and operate in various applications, it is seen an increase in malfunctioning of devices due to compromises in security aspects. An increase in the number of un-trustworthy activities leads to loosing of end users to any service provider. So all edge devices must be labeled as trustworthy or not, based on their previous transactions, leading to effective communications. Finding and maintaining the trust score of edge devices is the most pressing concern in the distributed communication environment. Considering all the issues, this paper propose a Machine Learning-based Trust Prediction for Edge Devices in the Blockchain enabled Fog Computing Environment (MLTPED-BFC). The proposed scheme uses an ensemble of Support Vector Regression (SVR) and Multivariable Logistic Regression (MLR) for predicting the trust score of each edge device and updates it after every successful communication. The prediction and updating of the trust score is carried out by the fog server without any biasing. This Artificial Intelligence driven approach enhances communication effectiveness and security by classifying devices as trustworthy or not, improving the overall reliability of the distributed system. The proposed scheme is proved to be secured based on informal security analysis. Extensive simulations are carried out to validate the proposed scheme's effectiveness and compare it with existing schemes. The proposed MLTPED-BFC mechanism have attained 98.91% of accuracy, 0.0048 loss rate, 98.92% of precision, 98.32% of recall, 98.96% of F-Measure and took 356 s for 100 iterations.
{"title":"MLTPED-BFC: Machine learning-based trust prediction for edge devices in the blockchain enabled fog computing environment","authors":"","doi":"10.1016/j.engappai.2024.109518","DOIUrl":"10.1016/j.engappai.2024.109518","url":null,"abstract":"<div><div>The utilization of edge devices in fog computing services is increasing every day to achieve effective communication between edge devices as it reduces the latency and processing time. When the number of edge devices increases and operate in various applications, it is seen an increase in malfunctioning of devices due to compromises in security aspects. An increase in the number of un-trustworthy activities leads to loosing of end users to any service provider. So all edge devices must be labeled as trustworthy or not, based on their previous transactions, leading to effective communications. Finding and maintaining the trust score of edge devices is the most pressing concern in the distributed communication environment. Considering all the issues, this paper propose a Machine Learning-based Trust Prediction for Edge Devices in the Blockchain enabled Fog Computing Environment (MLTPED-BFC). The proposed scheme uses an ensemble of Support Vector Regression (SVR) and Multivariable Logistic Regression (MLR) for predicting the trust score of each edge device and updates it after every successful communication. The prediction and updating of the trust score is carried out by the fog server without any biasing. This Artificial Intelligence driven approach enhances communication effectiveness and security by classifying devices as trustworthy or not, improving the overall reliability of the distributed system. The proposed scheme is proved to be secured based on informal security analysis. Extensive simulations are carried out to validate the proposed scheme's effectiveness and compare it with existing schemes. The proposed MLTPED-BFC mechanism have attained 98.91% of accuracy, 0.0048 loss rate, 98.92% of precision, 98.32% of recall, 98.96% of F-Measure and took 356 s for 100 iterations.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":null,"pages":null},"PeriodicalIF":7.5,"publicationDate":"2024-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142533412","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2024-10-28DOI: 10.1016/j.engappai.2024.109496
Video Anomaly Detection (VAD) for weakly supervised data operates with limited video-level annotations. It also holds the practical significance to play a pivotal role in surveillance and security applications like public safety, patient monitoring, autonomous vehicles, etc. Moreover, VAD extends its utility to various industrial settings, where it is instrumental in safeguarding workers' safety, enabling real-time production quality monitoring, and predictive maintenance. These diverse applications highlight the versatility of VAD and its potential to transform processes across various industries, making it an essential tool along with traditional surveillance applications. The majority of the existing studies have been focused on mitigating critical aspects of VAD, such as reducing false alarm rates and misdetection. These challenges can be effectively addressed by capturing the intricate spatiotemporal pattern within video data. Therefore, the proposed work named Swin Transformer-based Hybrid Temporal Adaptive Module (ST-HTAM) Abnormal Event Detection introduces an intuitive temporal module along with leveraging the strengths of the Swin (Shifted window-based) Transformers for spatial analysis. The novel aspect of this work lies in the hybridization of global self-attention and Convolutional-Long Short Term Memory (C-LSTM) Networks are renowned for capturing both global and local temporal dependencies. By extracting these spatial and temporal components, the proposed method, ST-HTAM, offers a comprehensive understanding of anomalous events. Altogether, it enhances the accuracy and robustness of Weakly Supervised VAD (WS-VAD). Finally, an anomaly scoring mechanism is employed in the classification step to facilitate effective anomaly detection from test video data. The proposed system is tailored to operate in real-time and highlights the dual focus on sophisticated Artificial Intelligence (AI) techniques and their impactful use cases across diverse domains. Comprehensive experiments are conducted on benchmark datasets that clearly show the substantial superiority of the ST-HTAM over state-of-the-art approaches. Code is available at https://github.com/Shalmiyapaulraj78/STHTAM-VAD.
针对弱监督数据的视频异常检测(VAD)可在有限的视频级注释下运行。它在公共安全、病人监控、自动驾驶汽车等监控和安全应用中发挥着举足轻重的作用。此外,VAD 还可应用于各种工业环境,在保障工人安全、实现实时生产质量监控和预测性维护方面发挥重要作用。这些多样化的应用凸显了 VAD 的多功能性及其改变各行业流程的潜力,使其成为传统监控应用的重要工具。现有的大部分研究都集中在减少 VAD 的关键方面,如降低误报率和错误检测。通过捕捉视频数据中错综复杂的时空模式,可以有效地应对这些挑战。因此,这项名为 "基于斯文变换器的混合时态自适应模块(ST-HTAM)异常事件检测 "的工作引入了一个直观的时态模块,并利用斯文(基于移位窗口的)变换器的优势进行空间分析。这项工作的新颖之处在于将全局自我注意与卷积-长短期记忆(C-LSTM)网络进行了混合,后者在捕捉全局和局部时间依赖性方面享有盛誉。通过提取这些空间和时间成分,所提出的 ST-HTAM 方法可以全面了解异常事件。总之,它提高了弱监督 VAD(WS-VAD)的准确性和鲁棒性。最后,在分类步骤中采用了异常评分机制,以促进从测试视频数据中进行有效的异常检测。所提出的系统是为实时运行而量身定制的,突出了对复杂的人工智能(AI)技术及其在不同领域的有影响力的用例的双重关注。在基准数据集上进行的综合实验清楚地表明,ST-HTAM 比最先进的方法更具实质性优势。代码见 https://github.com/Shalmiyapaulraj78/STHTAM-VAD。
{"title":"Transformer-enabled weakly supervised abnormal event detection in intelligent video surveillance systems","authors":"","doi":"10.1016/j.engappai.2024.109496","DOIUrl":"10.1016/j.engappai.2024.109496","url":null,"abstract":"<div><div>Video Anomaly Detection (VAD) for weakly supervised data operates with limited video-level annotations. It also holds the practical significance to play a pivotal role in surveillance and security applications like public safety, patient monitoring, autonomous vehicles, etc. Moreover, VAD extends its utility to various industrial settings, where it is instrumental in safeguarding workers' safety, enabling real-time production quality monitoring, and predictive maintenance. These diverse applications highlight the versatility of VAD and its potential to transform processes across various industries, making it an essential tool along with traditional surveillance applications. The majority of the existing studies have been focused on mitigating critical aspects of VAD, such as reducing false alarm rates and misdetection. These challenges can be effectively addressed by capturing the intricate spatiotemporal pattern within video data. Therefore, the proposed work named Swin Transformer-based Hybrid Temporal Adaptive Module (ST-HTAM) Abnormal Event Detection introduces an intuitive temporal module along with leveraging the strengths of the Swin (Shifted window-based) Transformers for spatial analysis. The novel aspect of this work lies in the hybridization of global self-attention and Convolutional-Long Short Term Memory (C-LSTM) Networks are renowned for capturing both global and local temporal dependencies. By extracting these spatial and temporal components, the proposed method, ST-HTAM, offers a comprehensive understanding of anomalous events. Altogether, it enhances the accuracy and robustness of Weakly Supervised VAD (WS-VAD). Finally, an anomaly scoring mechanism is employed in the classification step to facilitate effective anomaly detection from test video data. The proposed system is tailored to operate in real-time and highlights the dual focus on sophisticated Artificial Intelligence (AI) techniques and their impactful use cases across diverse domains. Comprehensive experiments are conducted on benchmark datasets that clearly show the substantial superiority of the ST-HTAM over state-of-the-art approaches. Code is available at <span><span>https://github.com/Shalmiyapaulraj78/STHTAM-VAD</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":null,"pages":null},"PeriodicalIF":7.5,"publicationDate":"2024-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142533411","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2024-10-28DOI: 10.1016/j.engappai.2024.109489
Real-time bearing fault diagnosis plays a vital role in maintaining the safety and reliability of sophisticated industrial systems. However, the scarcity of labeled data in fault diagnosis, due to the difficulty of collecting fault samples and the high cost of labeling, poses a significant challenge in learning discriminative fault features from limited and complex monitoring signals. Few-shot learning (FSL) emerges as a potent method for extracting and accurately classifying features from severe fault signals. Nonetheless, challenges such as data scarcity and environmental noise significantly impede the efficacy of existing FSL methods in diagnosing incipient faults effectively. These limitations are primarily due to the inadequate consideration of inter-class correlations within noisy contexts by current FSL strategies, which restricts their ability to extrapolate familiar features to new classes. Consequently, there is a pressing demand for an FSL approach that can exploit inter-class correlations to address the hurdles of data insufficiency and environmental complexities, thereby facilitating the diagnosis of incipient faults in few-shot settings. This paper proposes a novel category-knowledge-guided model tailored for few-shot multi-task scenarios. By leveraging attribute data from base categories and the similarities across new class samples, our model efficiently establishes mapping relations for unencountered tasks, significantly enhancing its generalization capabilities for early-stage fault diagnosis and multi-task applications. This model ensures swift and precise FSL fault diagnosis under uncharted operational conditions. Comparative analyses utilizing the Case Western Reserve University bearing dataset and the Early Mild Fault Traction Motor bearing dataset demonstrate our model’s superior performance against leading FSL and transfer learning approaches.
{"title":"Category knowledge-guided few-shot bearing fault diagnosis","authors":"","doi":"10.1016/j.engappai.2024.109489","DOIUrl":"10.1016/j.engappai.2024.109489","url":null,"abstract":"<div><div>Real-time bearing fault diagnosis plays a vital role in maintaining the safety and reliability of sophisticated industrial systems. However, the scarcity of labeled data in fault diagnosis, due to the difficulty of collecting fault samples and the high cost of labeling, poses a significant challenge in learning discriminative fault features from limited and complex monitoring signals. Few-shot learning (FSL) emerges as a potent method for extracting and accurately classifying features from severe fault signals. Nonetheless, challenges such as data scarcity and environmental noise significantly impede the efficacy of existing FSL methods in diagnosing incipient faults effectively. These limitations are primarily due to the inadequate consideration of inter-class correlations within noisy contexts by current FSL strategies, which restricts their ability to extrapolate familiar features to new classes. Consequently, there is a pressing demand for an FSL approach that can exploit inter-class correlations to address the hurdles of data insufficiency and environmental complexities, thereby facilitating the diagnosis of incipient faults in few-shot settings. This paper proposes a novel category-knowledge-guided model tailored for few-shot multi-task scenarios. By leveraging attribute data from base categories and the similarities across new class samples, our model efficiently establishes mapping relations for unencountered tasks, significantly enhancing its generalization capabilities for early-stage fault diagnosis and multi-task applications. This model ensures swift and precise FSL fault diagnosis under uncharted operational conditions. Comparative analyses utilizing the Case Western Reserve University bearing dataset and the Early Mild Fault Traction Motor bearing dataset demonstrate our model’s superior performance against leading FSL and transfer learning approaches.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":null,"pages":null},"PeriodicalIF":7.5,"publicationDate":"2024-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142533288","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2024-10-28DOI: 10.1016/j.engappai.2024.109498
This study focuses on applying the algorithm of Dueling Double Deep Q Networks to create a robust and adaptable navigation system for autonomous robots. The main objective of the study is to propose a network model that is capable of making a robot explore unknown areas, avoid static and dynamic obstacles efficiently, recognize predefined targets, and achieve them with high accuracy. Toward this, three different network models have been designed and trained with depth images obtained from a depth camera and directional and distance information from an RGB camera. First, these models were trained and tested in simple and complex simulated environments. The D3QN-C model demonstrated strong performance, achieving a success rate of 89% in the simple environment and of 87% in the complex environment. Tests were further extended by adding real-world data with different obstacle densities in order to prove the strength of the model in increasingly difficult and realistic scenarios. During all tests, the D3QN-C model could sustain high performance, showing 90% success rates in low-density, 85% in medium-density, and 82% in high-density environments. These results are evidence of the efficiency, robustness, and flexibility of this model and underline the potential of the algorithm of the Dueling Double Deep Q Networks as a main tool in using robots within real-world scenarios characterized by dynamics and complexity.
{"title":"Navigation of autonomous mobile robots in dynamic unknown environments based on dueling double deep q networks","authors":"","doi":"10.1016/j.engappai.2024.109498","DOIUrl":"10.1016/j.engappai.2024.109498","url":null,"abstract":"<div><div>This study focuses on applying the algorithm of Dueling Double Deep Q Networks to create a robust and adaptable navigation system for autonomous robots. The main objective of the study is to propose a network model that is capable of making a robot explore unknown areas, avoid static and dynamic obstacles efficiently, recognize predefined targets, and achieve them with high accuracy. Toward this, three different network models have been designed and trained with depth images obtained from a depth camera and directional and distance information from an RGB camera. First, these models were trained and tested in simple and complex simulated environments. The D3QN-C model demonstrated strong performance, achieving a success rate of 89% in the simple environment and of 87% in the complex environment. Tests were further extended by adding real-world data with different obstacle densities in order to prove the strength of the model in increasingly difficult and realistic scenarios. During all tests, the D3QN-C model could sustain high performance, showing 90% success rates in low-density, 85% in medium-density, and 82% in high-density environments. These results are evidence of the efficiency, robustness, and flexibility of this model and underline the potential of the algorithm of the Dueling Double Deep Q Networks as a main tool in using robots within real-world scenarios characterized by dynamics and complexity.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":null,"pages":null},"PeriodicalIF":7.5,"publicationDate":"2024-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142533413","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2024-10-28DOI: 10.1016/j.engappai.2024.109503
A grid-connected photovoltaic system integrates solar panels with the utility grid through a power inverter unit, allowing them to operate in parallel with the grid. Commonly known as grid-tied or on-grid solar systems, these configurations enable panels to feed electrical energy back into the grid, offering simplicity, low operating and maintenance costs, and reduced electricity bills. Despite these advantages, this environmentally friendly energy solution is still susceptible to downtimes and faults. This study utilizes advanced machine learning tree-based algorithms for fault detection and diagnosis in such systems with the goal of maintaining reliability, improving performance, and ensuring optimal energy generation. Specifically, the research investigates the effectiveness of Extra Trees as a fault detection and diagnosis algorithm through an efficient two-phase framework that consists of a binary fault detection phase followed by a multi-class fault diagnosis phase, achieving respective accuracies of 99.5% and 98.7%. In addition, the study underscores the importance of oversampling in improving results, particularly for imbalanced datasets. Moreover, explainable artificial intelligence is employed to enhance transparency in the model’s output and sensitivity to specific features in a given order. Remarkably, the findings align directly with results obtained from techniques such as feature importance averaging and incremental feature accuracy tracking. The research unveils a highly scalable, lightweight, and simple framework for fault detection and diagnosis in grid-connected photovoltaic systems.
并网光伏系统通过电力逆变器装置将太阳能电池板与公用电网整合在一起,使其能够与电网并网运行。这些配置通常被称为并网型或并网型太阳能系统,可使太阳能电池板将电能反馈给电网,从而提供简便性、低运行和维护成本,并减少电费支出。尽管有这些优点,但这种环保能源解决方案仍然容易出现停机和故障。本研究利用先进的基于机器学习树的算法对此类系统进行故障检测和诊断,目的是保持可靠性、提高性能并确保最佳发电效果。具体来说,该研究通过一个高效的两阶段框架(包括二进制故障检测阶段和多类故障诊断阶段),研究了 Extra Trees 作为故障检测和诊断算法的有效性,其准确率分别达到 99.5% 和 98.7%。此外,该研究还强调了超采样对改善结果的重要性,尤其是对不平衡数据集而言。此外,还采用了可解释人工智能,以提高模型输出的透明度和对特定顺序的特定特征的敏感性。值得注意的是,研究结果与特征重要性平均法和增量特征准确性跟踪等技术得出的结果直接吻合。这项研究为并网光伏系统的故障检测和诊断揭开了一个高度可扩展、轻量级和简单的框架。
{"title":"Explainable artificial intelligence of tree-based algorithms for fault detection and diagnosis in grid-connected photovoltaic systems","authors":"","doi":"10.1016/j.engappai.2024.109503","DOIUrl":"10.1016/j.engappai.2024.109503","url":null,"abstract":"<div><div>A grid-connected photovoltaic system integrates solar panels with the utility grid through a power inverter unit, allowing them to operate in parallel with the grid. Commonly known as grid-tied or on-grid solar systems, these configurations enable panels to feed electrical energy back into the grid, offering simplicity, low operating and maintenance costs, and reduced electricity bills. Despite these advantages, this environmentally friendly energy solution is still susceptible to downtimes and faults. This study utilizes advanced machine learning tree-based algorithms for fault detection and diagnosis in such systems with the goal of maintaining reliability, improving performance, and ensuring optimal energy generation. Specifically, the research investigates the effectiveness of Extra Trees as a fault detection and diagnosis algorithm through an efficient two-phase framework that consists of a binary fault detection phase followed by a multi-class fault diagnosis phase, achieving respective accuracies of 99.5% and 98.7%. In addition, the study underscores the importance of oversampling in improving results, particularly for imbalanced datasets. Moreover, explainable artificial intelligence is employed to enhance transparency in the model’s output and sensitivity to specific features in a given order. Remarkably, the findings align directly with results obtained from techniques such as feature importance averaging and incremental feature accuracy tracking. The research unveils a highly scalable, lightweight, and simple framework for fault detection and diagnosis in grid-connected photovoltaic systems.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":null,"pages":null},"PeriodicalIF":7.5,"publicationDate":"2024-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142533409","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2024-10-28DOI: 10.1016/j.engappai.2024.109509
Clustering is essential for uncovering hidden patterns and relationships in complex datasets. Its importance reveals when labeled data is scarce, expensive, time-consuming to obtain. Real-world applications often exhibit heterogeneity due to the diverse nature of the encapsulated data. This heterogeneity poses a significant challenge in data analysis, modeling, and makes traditional clustering methods ineffective. By adopting a hybrid architecture based on two promising techniques, multi-view and deep clustering, our method achieved better results, outperforming several existing methods including K-means, deep embedded clustering, deep clustering network, deep embedded K-means among many others. Multiple experiments conducted across diverse publicly accessible datasets validate the effectiveness of our proposed method based on well established evaluation metrics such as Accuracy and Normalized Mutual Information (NMI). Furthermore, we applied our method on the air pollution data of Luxembourg, a country with sparse sensor coverage. Our method demonstrated promising results, and unveil a new dimension that pave way for future work in air pollution’s level prediction and hotspots detection, crucial steps towards effective pollution reduction strategies.
{"title":"Multi-view Deep Embedded Clustering: Exploring a new dimension of air pollution","authors":"","doi":"10.1016/j.engappai.2024.109509","DOIUrl":"10.1016/j.engappai.2024.109509","url":null,"abstract":"<div><div>Clustering is essential for uncovering hidden patterns and relationships in complex datasets. Its importance reveals when labeled data is scarce, expensive, time-consuming to obtain. Real-world applications often exhibit heterogeneity due to the diverse nature of the encapsulated data. This heterogeneity poses a significant challenge in data analysis, modeling, and makes traditional clustering methods ineffective. By adopting a hybrid architecture based on two promising techniques, multi-view and deep clustering, our method achieved better results, outperforming several existing methods including <em>K</em>-means, deep embedded clustering, deep clustering network, deep embedded <em>K</em>-means among many others. Multiple experiments conducted across diverse publicly accessible datasets validate the effectiveness of our proposed method based on well established evaluation metrics such as Accuracy and Normalized Mutual Information (NMI). Furthermore, we applied our method on the air pollution data of Luxembourg, a country with sparse sensor coverage. Our method demonstrated promising results, and unveil a new dimension that pave way for future work in air pollution’s level prediction and hotspots detection, crucial steps towards effective pollution reduction strategies.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":null,"pages":null},"PeriodicalIF":7.5,"publicationDate":"2024-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142533156","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2024-10-28DOI: 10.1016/j.engappai.2024.109527
Code smells are software flaws that make it challenging to comprehend, develop, and maintain the software. Identifying and removing code smells is crucial for software quality. This study examines the effectiveness of several machine-learning models before and after applying feature selection and data balancing on code smell datasets. Extreme Gradient Boosting, Gradient Boosting, Adaptive Boosting, Random Forest, Artificial Neural Network (ANN), and Ensemble model of Bagging, and the two best-performing Boosting techniques are used to predict code smell. This study proposes an enhanced approach, which is an ensemble model of the Bagging and Boosting classifier (EMBBC) that incorporates feature selection and data balancing techniques to predict code smells. Four publicly available code smell datasets, Blob Class, Data Class, Long Parameter List, and Switch Statement, were considered for the experimental work. Classes of datasets are balanced using the Synthetic Minority Over-Sampling Technique (SMOTE). A feature selection method called Recursive Feature Elimination with Cross-Validation (RFECV) is used. This study shows that the ensemble model of Bagging and the two best-performing Boosting techniques performs better in Blob Class, Data Class, and Long Parameter List datasets with the highest accuracy of 99.21%, 99.21%, and 97.62%, respectively. In the Switch Statement dataset, the ANN model provides a higher accuracy of 92.86%. Since the proposed model uses only seven features and still provides better results than others, it could be helpful to detect code smells for software engineers and practitioners in less computational time, improving the system's overall performance.
{"title":"Ensemble methods with feature selection and data balancing for improved code smells classification performance","authors":"","doi":"10.1016/j.engappai.2024.109527","DOIUrl":"10.1016/j.engappai.2024.109527","url":null,"abstract":"<div><div>Code smells are software flaws that make it challenging to comprehend, develop, and maintain the software. Identifying and removing code smells is crucial for software quality. This study examines the effectiveness of several machine-learning models before and after applying feature selection and data balancing on code smell datasets. Extreme Gradient Boosting, Gradient Boosting, Adaptive Boosting, Random Forest, Artificial Neural Network (ANN), and Ensemble model of Bagging, and the two best-performing Boosting techniques are used to predict code smell. This study proposes an enhanced approach, which is an ensemble model of the Bagging and Boosting classifier (EMBBC) that incorporates feature selection and data balancing techniques to predict code smells. Four publicly available code smell datasets, Blob Class, Data Class, Long Parameter List, and Switch Statement, were considered for the experimental work. Classes of datasets are balanced using the Synthetic Minority Over-Sampling Technique (SMOTE). A feature selection method called Recursive Feature Elimination with Cross-Validation (RFECV) is used. This study shows that the ensemble model of Bagging and the two best-performing Boosting techniques performs better in Blob Class, Data Class, and Long Parameter List datasets with the highest accuracy of 99.21%, 99.21%, and 97.62%, respectively. In the Switch Statement dataset, the ANN model provides a higher accuracy of 92.86%. Since the proposed model uses only seven features and still provides better results than others, it could be helpful to detect code smells for software engineers and practitioners in less computational time, improving the system's overall performance.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":null,"pages":null},"PeriodicalIF":7.5,"publicationDate":"2024-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142533410","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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}