Engineering Applications of Artificial Intelligence最新文献_第4页

A modified multi-agent proximal policy optimization algorithm for multi-objective dynamic partial-re-entrant hybrid flow shop scheduling problem 针对多目标动态部分重入混合流车间调度问题的改进型多代理近端策略优化算法

IF 7.5 2区计算机科学 Q1 AUTOMATION & CONTROL SYSTEMS

Engineering Applications of Artificial Intelligence

Pub Date : 2024-11-27 DOI: 10.1016/j.engappai.2024.109688

Jiawei Wu, Yong Liu

This paper extends a novel model for modern flexible manufacturing systems: the multi-objective dynamic partial-re-entrant hybrid flow shop scheduling problem (MDPR-HFSP). The model considers partial-re-entrant processing, dynamic disturbance events, green manufacturing demand, and machine workload. Despite advancements in applying deep reinforcement learning to dynamic workshop scheduling, current methods face challenges in training scheduling policies for partial-re-entrant processing constraints and multiple manufacturing objectives. To solve the MDPR-HFSP, we propose a modified multi-agent proximal policy optimization (MMAPPO) algorithm, which employs a routing agent (RA) for machine assignment and a sequencing agent (SA) for job selection. Four machine assignment rules and four job selection rules are integrated to choose optimum actions for RA and SA at rescheduling points. In addition, reward signals are created by combining objective weight vectors with reward vectors, and training parameters under each weight vector are saved to flexibly optimize three objectives. Furthermore, we design an adaptive trust region clipping method to improve the constraint of the proximal policy optimization algorithm on the differences between new and old policies by introducing the Wasserstein distance. Moreover, we conduct comprehensive numerical experiments to compare the proposed MMAPPO algorithm with nine composite scheduling rules and the basic multi-agent proximal policy optimization algorithm. The results demonstrate that the proposed MMAPPO algorithm is more effective in solving the MDPR-HFSP and achieves superior convergence and diversity in solutions. Finally, a semiconductor wafer manufacturing case is resolved by the MMAPPO, and the scheduling solution meets the responsive requirement.

本文扩展了现代柔性制造系统的一个新模型：多目标动态部分再入站混合流程车间调度问题（MDPR-HFSP）。该模型考虑了部分重入加工、动态干扰事件、绿色制造需求和机器工作量。尽管在将深度强化学习应用于动态车间调度方面取得了进展，但目前的方法在针对部分再入加工约束和多重制造目标训练调度策略方面仍面临挑战。为解决 MDPR-HFSP 问题，我们提出了一种改进的多代理近端策略优化（MMAPPO）算法，该算法采用路由代理（RA）进行机器分配，采用排序代理（SA）进行作业选择。四种机器分配规则和四种作业选择规则被整合在一起，为 RA 和 SA 在重新安排点选择最佳行动。此外，通过将目标权重向量与奖励向量相结合来创建奖励信号，并保存每个权重向量下的训练参数，从而灵活优化三个目标。此外，我们还设计了一种自适应信任区域剪切方法，通过引入瓦瑟斯坦距离来改善近似策略优化算法对新旧策略差异的约束。此外，我们还进行了全面的数值实验，将提出的 MMAPPO 算法与九种复合调度规则和基本的多代理近端策略优化算法进行了比较。结果表明，所提出的 MMAPPO 算法在求解 MDPR-HFSP 时更为有效，并实现了更优越的收敛性和解的多样性。最后，MMAPPO 解决了一个半导体晶圆制造案例，其调度方案符合响应要求。

{"title":"A modified multi-agent proximal policy optimization algorithm for multi-objective dynamic partial-re-entrant hybrid flow shop scheduling problem","authors":"Jiawei Wu, Yong Liu","doi":"10.1016/j.engappai.2024.109688","DOIUrl":"10.1016/j.engappai.2024.109688","url":null,"abstract":"<div><div>This paper extends a novel model for modern flexible manufacturing systems: the multi-objective dynamic partial-re-entrant hybrid flow shop scheduling problem (MDPR-HFSP). The model considers partial-re-entrant processing, dynamic disturbance events, green manufacturing demand, and machine workload. Despite advancements in applying deep reinforcement learning to dynamic workshop scheduling, current methods face challenges in training scheduling policies for partial-re-entrant processing constraints and multiple manufacturing objectives. To solve the MDPR-HFSP, we propose a modified multi-agent proximal policy optimization (MMAPPO) algorithm, which employs a routing agent (RA) for machine assignment and a sequencing agent (SA) for job selection. Four machine assignment rules and four job selection rules are integrated to choose optimum actions for RA and SA at rescheduling points. In addition, reward signals are created by combining objective weight vectors with reward vectors, and training parameters under each weight vector are saved to flexibly optimize three objectives. Furthermore, we design an adaptive trust region clipping method to improve the constraint of the proximal policy optimization algorithm on the differences between new and old policies by introducing the Wasserstein distance. Moreover, we conduct comprehensive numerical experiments to compare the proposed MMAPPO algorithm with nine composite scheduling rules and the basic multi-agent proximal policy optimization algorithm. The results demonstrate that the proposed MMAPPO algorithm is more effective in solving the MDPR-HFSP and achieves superior convergence and diversity in solutions. Finally, a semiconductor wafer manufacturing case is resolved by the MMAPPO, and the scheduling solution meets the responsive requirement.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"140 ","pages":"Article 109688"},"PeriodicalIF":7.5,"publicationDate":"2024-11-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142723295","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}

引用次数: 0

LCRTR-Net: A low-cost real-time recognition network for rail corrugation in railway transportation LCRTR-Net：用于铁路运输中轨道波纹的低成本实时识别网络

IF 7.5 2区计算机科学 Q1 AUTOMATION & CONTROL SYSTEMS

Engineering Applications of Artificial Intelligence

Pub Date : 2024-11-26 DOI: 10.1016/j.engappai.2024.109708

Xueyang Tang , Xiaopei Cai , Yuqi Wang , Yue Hou

Rail corrugation has a significant impact on the safety of high-speed railway operations, making its identification particularly important. Traditional manual inspection methods are infeasible for large-scale identification within limited time frames, while existing methods based on machine vision or axle box acceleration face challenges such as high costs, complex equipment installation and maintenance, as well as difficulties in achieving real-time performance. To address these challenges, this study proposes an innovative low-cost real-time recognition network (LCRTR-Net), which utilizes accelerometers installed on the underside of the train body and combines wavelet packet decomposition with dilated causal convolution in a residual neural network. Specifically, the approach first extracts the latent features of train body acceleration caused by rail corrugation through wavelet packet decomposition and reconstruction. Next, dilated causal convolution is employed to capture the temporal causal relationships and long-term dependencies of these latent features. Finally, the integration of residual connections further enhances the feature extraction performance and computational efficiency of LCRTR-Net. Experimental results demonstrate that LCRTR-Net exhibits significant generalization ability and real-time performance, achieving an average recognition accuracy exceeding 97.0%, with a recognition time of only 0.17 ms per rail corrugation sample, significantly outperforming existing rail corrugation recognition methods. This indicates that LCRTR-Net has broad application potential in practical railway operations. Future research directions will focus on unsupervised or few-shot learning algorithms and multi-sensor integration to further improve recognition accuracy and real-time performance, promoting the practical application of this technology.

轨道波纹对高速铁路运营的安全性有重大影响，因此对其进行识别尤为重要。传统的人工检测方法无法在有限的时间内进行大规模识别，而现有的基于机器视觉或轴箱加速度的方法也面临着成本高、设备安装和维护复杂以及难以实现实时性能等挑战。为了应对这些挑战，本研究提出了一种创新的低成本实时识别网络（LCRTR-Net），它利用安装在列车车身底部的加速度计，并在残差神经网络中结合了小波包分解和扩张因果卷积。具体来说，该方法首先通过小波包分解和重构提取轨道波纹引起的列车车身加速度的潜在特征。然后，利用扩张因果卷积来捕捉这些潜特征的时间因果关系和长期依赖关系。最后，残差连接的整合进一步提高了 LCRTR-Net 的特征提取性能和计算效率。实验结果表明，LCRTR-Net 具有显著的泛化能力和实时性，平均识别准确率超过 97.0%，每个铁路波纹样本的识别时间仅为 0.17 毫秒，明显优于现有的铁路波纹识别方法。这表明 LCRTR-Net 在实际铁路运营中具有广泛的应用潜力。未来的研究方向将集中在无监督或少量学习算法以及多传感器集成方面，以进一步提高识别精度和实时性，促进该技术的实际应用。

{"title":"LCRTR-Net: A low-cost real-time recognition network for rail corrugation in railway transportation","authors":"Xueyang Tang , Xiaopei Cai , Yuqi Wang , Yue Hou","doi":"10.1016/j.engappai.2024.109708","DOIUrl":"10.1016/j.engappai.2024.109708","url":null,"abstract":"<div><div>Rail corrugation has a significant impact on the safety of high-speed railway operations, making its identification particularly important. Traditional manual inspection methods are infeasible for large-scale identification within limited time frames, while existing methods based on machine vision or axle box acceleration face challenges such as high costs, complex equipment installation and maintenance, as well as difficulties in achieving real-time performance. To address these challenges, this study proposes an innovative low-cost real-time recognition network (LCRTR-Net), which utilizes accelerometers installed on the underside of the train body and combines wavelet packet decomposition with dilated causal convolution in a residual neural network. Specifically, the approach first extracts the latent features of train body acceleration caused by rail corrugation through wavelet packet decomposition and reconstruction. Next, dilated causal convolution is employed to capture the temporal causal relationships and long-term dependencies of these latent features. Finally, the integration of residual connections further enhances the feature extraction performance and computational efficiency of LCRTR-Net. Experimental results demonstrate that LCRTR-Net exhibits significant generalization ability and real-time performance, achieving an average recognition accuracy exceeding 97.0%, with a recognition time of only 0.17 ms per rail corrugation sample, significantly outperforming existing rail corrugation recognition methods. This indicates that LCRTR-Net has broad application potential in practical railway operations. Future research directions will focus on unsupervised or few-shot learning algorithms and multi-sensor integration to further improve recognition accuracy and real-time performance, promoting the practical application of this technology.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"140 ","pages":"Article 109708"},"PeriodicalIF":7.5,"publicationDate":"2024-11-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142700728","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}

引用次数: 0

Dual-branch feature Reinforcement Transformer for preoperative parathyroid gland segmentation 用于术前甲状旁腺分割的双分支特征增强变换器

IF 7.5 2区计算机科学 Q1 AUTOMATION & CONTROL SYSTEMS

Engineering Applications of Artificial Intelligence

Pub Date : 2024-11-26 DOI: 10.1016/j.engappai.2024.109672

Lei Lyu , Chen Pang , Qinghan Yang , Kailin Liu , Chong Geng

The parathyroid glands are easily injured or accidentally removed during thyroid surgery, causing temporary or even permanent hypocalcemia. Thus, accurate preoperative identification and localization of the parathyroid glands by ultrasound is crucial in protecting the parathyroid glands and preventing parathyroid injury during thyroid surgery. However, there are only a few methods used for highlighting the parathyroid gland in ultrasound images before thyroid surgery. In this study, we propose a Dual-branch feature Reinforcement Transformer Network (DRT-Net) for preoperative parathyroid gland segmentation. DRT-Net incorporates a dual-branch structure, consisting of a devised convolution network (CNN) backbone called Feature Reinforcement subnet (FR-subnet) and a Transformer branch capturing detailed features and context information from the confused ultrasound image. Furthermore, we design a Margin Tracking Attention (MTA) that optimizes the ability of FR-subnet to process margin information by tracking margin pixels of feature map. Finally, we employ a Cross-channel Feature Reinforcement Module (CFRM) to fuse the extracted detailed features from the CNN branch with the global context information from the Transformer branch. We trained and evaluated the DRT-Net on the self-built parathyroid gland segmentation dataset and an open-access Kvasir-SEG dataset. Extensive experiments have been carried out to validate the efficiency of our method.

甲状腺手术中很容易损伤或意外切除甲状旁腺，从而导致暂时性甚至永久性低钙血症。因此，术前通过超声准确识别和定位甲状旁腺对于保护甲状旁腺和防止甲状腺手术中的甲状旁腺损伤至关重要。然而，目前只有少数几种方法能在甲状腺手术前通过超声图像突出显示甲状旁腺。在这项研究中，我们提出了一种用于术前甲状旁腺分割的双分支特征增强变换网络（DRT-Net）。DRT-Net 采用双分支结构，包括一个称为特征增强子网（FR-subnet）的设计卷积网络（CNN）主干和一个从混淆的超声图像中捕捉详细特征和上下文信息的变换器分支。此外，我们还设计了边缘跟踪注意（MTA），通过跟踪特征图的边缘像素来优化 FR 子网处理边缘信息的能力。最后，我们采用了跨通道特征增强模块（CFRM），将从 CNN 分支提取的细节特征与从 Transformer 分支提取的全局上下文信息进行融合。我们在自建的甲状旁腺分割数据集和开放访问的 Kvasir-SEG 数据集上对 DRT-Net 进行了训练和评估。为了验证我们方法的效率，我们进行了广泛的实验。

{"title":"Dual-branch feature Reinforcement Transformer for preoperative parathyroid gland segmentation","authors":"Lei Lyu , Chen Pang , Qinghan Yang , Kailin Liu , Chong Geng","doi":"10.1016/j.engappai.2024.109672","DOIUrl":"10.1016/j.engappai.2024.109672","url":null,"abstract":"<div><div>The parathyroid glands are easily injured or accidentally removed during thyroid surgery, causing temporary or even permanent hypocalcemia. Thus, accurate preoperative identification and localization of the parathyroid glands by ultrasound is crucial in protecting the parathyroid glands and preventing parathyroid injury during thyroid surgery. However, there are only a few methods used for highlighting the parathyroid gland in ultrasound images before thyroid surgery. In this study, we propose a Dual-branch feature Reinforcement Transformer Network (DRT-Net) for preoperative parathyroid gland segmentation. DRT-Net incorporates a dual-branch structure, consisting of a devised convolution network (CNN) backbone called Feature Reinforcement subnet (FR-subnet) and a Transformer branch capturing detailed features and context information from the confused ultrasound image. Furthermore, we design a Margin Tracking Attention (MTA) that optimizes the ability of FR-subnet to process margin information by tracking margin pixels of feature map. Finally, we employ a Cross-channel Feature Reinforcement Module (CFRM) to fuse the extracted detailed features from the CNN branch with the global context information from the Transformer branch. We trained and evaluated the DRT-Net on the self-built parathyroid gland segmentation dataset and an open-access Kvasir-SEG dataset. Extensive experiments have been carried out to validate the efficiency of our method.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"140 ","pages":"Article 109672"},"PeriodicalIF":7.5,"publicationDate":"2024-11-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142700717","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}

引用次数: 0

Deep reinforcement learning optimizer based novel Caputo fractional order sliding mode data driven controller 基于深度强化学习优化器的新型卡普托分数阶滑动模式数据驱动控制器

IF 7.5 2区计算机科学 Q1 AUTOMATION & CONTROL SYSTEMS

Engineering Applications of Artificial Intelligence

Pub Date : 2024-11-25 DOI: 10.1016/j.engappai.2024.109725

Amir Veisi , Hadi Delavari

The design of controllers in engineering applications typically requires a model that accurately captures the dynamics of the real system. However, finding a precise model for controller design can be challenging in real engineering applications. Consequently, data-driven methods have gained widespread use in engineering systems. This paper presents a novel robust data-driven fractional-order controller optimized through deep reinforcement learning. Additionally, a new robust fractional-order observer has been introduced to improve both the robustness and speed of the system. To establish the stability of the proposed control system, a new Lyapunov stability theorem based on the Caputo fractional-order definition is provided. The proposed controller offers significant advantages, including enhanced robustness against external disturbances, increased resilience to parameter uncertainties and unmodeled nonlinear dynamics, improved accuracy, greater speed, and guaranteed optimal control coefficients. Furthermore, assured adaptability is demonstrated due to the optimization provided by deep reinforcement learning including enhanced robustness against external disturbances, uncertainties of parameters, and unmodeled nonlinear dynamics; improved accuracy; greater speed; and guaranteed optimal control coefficients. Furthermore, assured adaptability is demonstrated due to the optimization provided by deep reinforcement learning. The performance of the proposed method has been compared with that of conventional integer-order sliding mode control, highlighting the superiority of this approach. The proposed method has been evaluated under normal conditions, external disturbances, and system uncertainties. Notably, performance improvements of 15%, 30%, and 68% have been achieved under normal conditions, external disturbances, and internal uncertainties, respectively, compared to the conventional integer-order sliding mode controller.

工程应用中的控制器设计通常需要一个能准确捕捉真实系统动态的模型。然而，在实际工程应用中，找到用于控制器设计的精确模型可能具有挑战性。因此，数据驱动方法在工程系统中得到了广泛应用。本文介绍了一种通过深度强化学习优化的新型鲁棒数据驱动分数阶控制器。此外，还引入了一种新的鲁棒分数阶观测器，以提高系统的鲁棒性和速度。为了确定所提控制系统的稳定性，提供了基于 Caputo 分数阶定义的新 Lyapunov 稳定性定理。所提出的控制器具有显著的优势，包括增强了对外部干扰的鲁棒性、提高了对参数不确定性和未建模非线性动态的适应能力、提高了精度、提高了速度并保证了最佳控制系数。此外，由于深度强化学习提供了优化功能，包括增强了对外界干扰、参数不确定性和未建模非线性动态的鲁棒性；提高了精度；提高了速度；以及保证了最优控制系数，因此可以确保适应性。此外，由于深度强化学习提供了优化功能，确保了适应性。所提方法的性能与传统的整数阶滑动模式控制进行了比较，凸显了该方法的优越性。在正常条件、外部干扰和系统不确定性条件下，对所提出的方法进行了评估。值得注意的是，与传统的整数阶滑动模式控制器相比，该方法在正常情况、外部干扰和内部不确定性下的性能分别提高了 15%、30% 和 68%。

{"title":"Deep reinforcement learning optimizer based novel Caputo fractional order sliding mode data driven controller","authors":"Amir Veisi , Hadi Delavari","doi":"10.1016/j.engappai.2024.109725","DOIUrl":"10.1016/j.engappai.2024.109725","url":null,"abstract":"<div><div>The design of controllers in engineering applications typically requires a model that accurately captures the dynamics of the real system. However, finding a precise model for controller design can be challenging in real engineering applications. Consequently, data-driven methods have gained widespread use in engineering systems. This paper presents a novel robust data-driven fractional-order controller optimized through deep reinforcement learning. Additionally, a new robust fractional-order observer has been introduced to improve both the robustness and speed of the system. To establish the stability of the proposed control system, a new Lyapunov stability theorem based on the Caputo fractional-order definition is provided. The proposed controller offers significant advantages, including enhanced robustness against external disturbances, increased resilience to parameter uncertainties and unmodeled nonlinear dynamics, improved accuracy, greater speed, and guaranteed optimal control coefficients. Furthermore, assured adaptability is demonstrated due to the optimization provided by deep reinforcement learning including enhanced robustness against external disturbances, uncertainties of parameters, and unmodeled nonlinear dynamics; improved accuracy; greater speed; and guaranteed optimal control coefficients. Furthermore, assured adaptability is demonstrated due to the optimization provided by deep reinforcement learning. The performance of the proposed method has been compared with that of conventional integer-order sliding mode control, highlighting the superiority of this approach. The proposed method has been evaluated under normal conditions, external disturbances, and system uncertainties. Notably, performance improvements of 15%, 30%, and 68% have been achieved under normal conditions, external disturbances, and internal uncertainties, respectively, compared to the conventional integer-order sliding mode controller.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"140 ","pages":"Article 109725"},"PeriodicalIF":7.5,"publicationDate":"2024-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142700719","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}

引用次数: 0

Assessing the factors of blockchain technology-enabled hospitals using an integrated interval-valued q-rung orthopair fuzzy decision-making model 利用综合区间值q-rung正交模糊决策模型评估区块链技术赋能医院的因素

IF 7.5 2区计算机科学 Q1 AUTOMATION & CONTROL SYSTEMS

Engineering Applications of Artificial Intelligence

Pub Date : 2024-11-25 DOI: 10.1016/j.engappai.2024.109641

Rashmi Pathak , Badal Soni , Naresh Babu Muppalaneni , Muhammet Deveci

In the current era, blockchain technology (BT) has emerged as a novel technique to maintain the operations of healthcare management systems. Assessment of blockchain technology (BT)-enabled hospitals can be considered as a multi-criteria decision making (MCDM) problem because of the existence of several criteria. The aim of this study is to develop a hybrid MCDM method for evaluating the factors of multi-criteria BT-enabled hospital selection problem under interval-valued q-rung orthopair fuzzy sets (IVq-ROFSs). For this purpose, a weighted aggregated sum product assessment (WASPAS) model is presented with the combination of IVq-ROF interaction aggregation operators, the standard deviation (SD)-based model and pivot pairwise relative criteria importance assessment (PIPRECIA) tool called IV-q-ROF-SD-PIPRECIA-WASPAS model within the context of IVq-ROFSs. For this purpose, some new IVq-ROF interaction aggregation operators are developed with their desirable characteristics. Next, the standard deviation-based model and PIPRECIA model on IVq-ROFSs are proposed to obtain the final weight of criteria, whereas the rank-based formula is presented to determine the decision experts’ weights with IVq-ROF information. The presented IV-q-ROF-SD-PIPRECIA-WASPAS model is applied on a case study of BT-enabled hospitals assessment, which confirms its applicability and usefulness. Sensitivity analysis and comparative discussion have been performed to reveal the consistency, robustness and efficiency of the presented model. The BT-enabled hospital-II with highest UD (0.4453) has emerged as the best choice among a set of BT-enabled hospitals. The factor "flexibilty" with highest weight (0.0898) value followed that the scalability (0.0809), transaction speed and accountability with same weight (0.0779) value, and network availability with weight (0.0771) for BT-enabled hospitals assessment. The final results conclude that the developed methodology can provide more accurate decisions while considering multiple indicators and input uncertainties.

在当今时代，区块链技术（BT）已成为维护医疗管理系统运行的一项新技术。由于存在多个标准，因此对支持区块链技术（BT）的医院进行评估可被视为一个多标准决策（MCDM）问题。本研究的目的是开发一种混合 MCDM 方法，用于评估区间值 q-rung 正交模糊集（IVq-ROFSs）下多标准 BT 支持的医院选择问题的因素。为此，结合 IVq-ROF 交互聚合算子、基于标准偏差（SD）的模型和枢轴成对相对标准重要性评估（PIPRECIA）工具，提出了一个加权聚合和积评估（WASPAS）模型，称为 IVq-ROFSs 下的 IV-q-ROF-SD-PIPRECIA-WASPAS 模型。为此，我们开发了一些新的 IVq-ROF 交互聚合算子，它们具有理想的特性。接下来，提出了基于标准偏差的模型和基于 IVq-ROFSs 的 PIPRECIA 模型来获得标准的最终权重，并提出了基于等级的公式来确定具有 IVq-ROF 信息的决策专家权重。将所提出的 IV-q-ROF-SD-PIPRECIA-WASPAS 模型应用于 BT 医院评估案例研究，证实了该模型的适用性和实用性。敏感性分析和比较讨论揭示了所提出模型的一致性、稳健性和效率。在一系列 BT 能力医院中，UD 最高（0.4453）的 BT 能力医院-II 成为最佳选择。灵活性 "因素的权重（0.0898）最高，其次是可扩展性（0.0809）、交易速度和责任（权重（0.0779）相同）以及网络可用性（权重（0.0771））。最终结果表明，所开发的方法可以在考虑多个指标和输入不确定性的同时提供更准确的决策。

{"title":"Assessing the factors of blockchain technology-enabled hospitals using an integrated interval-valued q-rung orthopair fuzzy decision-making model","authors":"Rashmi Pathak , Badal Soni , Naresh Babu Muppalaneni , Muhammet Deveci","doi":"10.1016/j.engappai.2024.109641","DOIUrl":"10.1016/j.engappai.2024.109641","url":null,"abstract":"<div><div>In the current era, blockchain technology (BT) has emerged as a novel technique to maintain the operations of healthcare management systems. Assessment of blockchain technology (BT)-enabled hospitals can be considered as a multi-criteria decision making (MCDM) problem because of the existence of several criteria. The aim of this study is to develop a hybrid MCDM method for evaluating the factors of multi-criteria BT-enabled hospital selection problem under interval-valued q-rung orthopair fuzzy sets (IVq-ROFSs). For this purpose, a weighted aggregated sum product assessment (WASPAS) model is presented with the combination of IVq-ROF interaction aggregation operators, the standard deviation (SD)-based model and pivot pairwise relative criteria importance assessment (PIPRECIA) tool called IV-q-ROF-SD-PIPRECIA-WASPAS model within the context of IVq-ROFSs. For this purpose, some new IVq-ROF interaction aggregation operators are developed with their desirable characteristics. Next, the standard deviation-based model and PIPRECIA model on IVq-ROFSs are proposed to obtain the final weight of criteria, whereas the rank-based formula is presented to determine the decision experts’ weights with IVq-ROF information. The presented IV-q-ROF-SD-PIPRECIA-WASPAS model is applied on a case study of BT-enabled hospitals assessment, which confirms its applicability and usefulness. Sensitivity analysis and comparative discussion have been performed to reveal the consistency, robustness and efficiency of the presented model. The BT-enabled hospital-II with highest UD (0.4453) has emerged as the best choice among a set of BT-enabled hospitals. The factor \"flexibilty\" with highest weight (0.0898) value followed that the scalability (0.0809), transaction speed and accountability with same weight (0.0779) value, and network availability with weight (0.0771) for BT-enabled hospitals assessment. The final results conclude that the developed methodology can provide more accurate decisions while considering multiple indicators and input uncertainties.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"139 ","pages":"Article 109641"},"PeriodicalIF":7.5,"publicationDate":"2024-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142720895","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}

引用次数: 0

WA-Net: Wavelet Integrated Attention Network for Silk and Bamboo character recognition WA-Net：用于丝竹字符识别的小波综合注意力网络

IF 7.5 2区计算机科学 Q1 AUTOMATION & CONTROL SYSTEMS

Engineering Applications of Artificial Intelligence

Pub Date : 2024-11-25 DOI: 10.1016/j.engappai.2024.109674

Shengnan Li, Chi Zhou, Kaili Wang

Chu Bamboo and Silk ancient Chinese character (CBSC) was originated in the Chu state over 2000 years ago, representing an intermediate script between oracle bone script and seal script. Existing text images have degraded and suffered damage due to their ancient historical origins and insufficient preservation. Due to distinct structural and stroke texture characteristics, significant differences exist between CBSC and contemporary characters, posing challenges for intelligent recognition. Targeting these aforementioned characteristics, we propose a method called Wavelet Integrated Attention Network (WA-Net). This method integrates discrete wavelet transform and attention mechanisms to extract more discriminative features from severe noise interference and degraded text images. Additionally, a dataset named Chu Bamboo and Silk 730 (Chu730) for CBSC recognition has been created due to the lack of publicly available datasets. WA-Net introduces the discrete wavelet attention among layer (L-DWT) to broaden the feature learning space of convolutional neural networks into the wavelet domain, capturing latent information across various frequencies. Subsequently, a wavelet convolution (C-DWT) module is proposed to mitigate the partial information loss of conventional convolution operations. In the W-bneck module, the SE (Squeeze-and-Excitation) attention module and average pooling downsampling are introduced to enhance the extraction of valuable feature maps. Extensive experiments were conducted, including a baseline method that achieved top-1 recognition accuracy of 87.42%. The proposed method achieved an accuracy of 89.27%, and other top-n results also significantly surpassed the baseline accuracy. Other experiment results demonstrate the superiority of the proposed modules and theirvaluable applications in ancient text intelligent recognition and cultural heritage digital preservation. Furthermore, this approach holds significant promise in facilitating the study of other handwritten or ancient characters recognition. Dataset and code are available at: https://github.com/Nancy45-ui/WA-Net.

楚竹帛古文字（CBSC）起源于 2000 多年前的楚国，是介于甲骨文和篆书之间的一种文字。现有的文字图像由于其古老的历史渊源和保存不足而退化和损坏。由于结构和笔画纹理特征不同，中国古代文字与现代文字之间存在显著差异，给智能识别带来了挑战。针对上述特点，我们提出了一种名为小波综合注意力网络（WA-Net）的方法。该方法整合了离散小波变换和注意力机制，可从严重的噪声干扰和劣化的文本图像中提取更多的识别特征。此外，由于缺乏公开可用的数据集，还创建了一个用于识别 CBSC 的数据集，名为 "楚竹丝 730"（Chu730）。WA-Net 引入了离散小波注意层（L-DWT），将卷积神经网络的特征学习空间扩展到小波域，捕捉各种频率的潜在信息。随后，还提出了小波卷积（C-DWT）模块，以减轻传统卷积操作的部分信息损失。在 W-bneck 模块中，引入了 SE（挤压-激发）注意模块和平均池化下采样，以增强对有价值特征图的提取。我们进行了广泛的实验，其中基线方法的识别准确率达到了 87.42%。所提出的方法达到了 89.27% 的准确率，其他 top-n 结果也大大超过了基准准确率。其他实验结果证明了所提模块的优越性及其在古文字智能识别和文化遗产数字保护方面的宝贵应用。此外，这种方法在促进其他手写或古文字识别研究方面也大有可为。数据集和代码见：https://github.com/Nancy45-ui/WA-Net。

{"title":"WA-Net: Wavelet Integrated Attention Network for Silk and Bamboo character recognition","authors":"Shengnan Li, Chi Zhou, Kaili Wang","doi":"10.1016/j.engappai.2024.109674","DOIUrl":"10.1016/j.engappai.2024.109674","url":null,"abstract":"<div><div>Chu Bamboo and Silk ancient Chinese character (CBSC) was originated in the Chu state over 2000 years ago, representing an intermediate script between oracle bone script and seal script. Existing text images have degraded and suffered damage due to their ancient historical origins and insufficient preservation. Due to distinct structural and stroke texture characteristics, significant differences exist between CBSC and contemporary characters, posing challenges for intelligent recognition. Targeting these aforementioned characteristics, we propose a method called Wavelet Integrated Attention Network (WA-Net). This method integrates discrete wavelet transform and attention mechanisms to extract more discriminative features from severe noise interference and degraded text images. Additionally, a dataset named Chu Bamboo and Silk 730 (Chu730) for CBSC recognition has been created due to the lack of publicly available datasets. WA-Net introduces the discrete wavelet attention among layer (L-DWT) to broaden the feature learning space of convolutional neural networks into the wavelet domain, capturing latent information across various frequencies. Subsequently, a wavelet convolution (C-DWT) module is proposed to mitigate the partial information loss of conventional convolution operations. In the W-bneck module, the SE (Squeeze-and-Excitation) attention module and average pooling downsampling are introduced to enhance the extraction of valuable feature maps. Extensive experiments were conducted, including a baseline method that achieved top-1 recognition accuracy of 87.42%. The proposed method achieved an accuracy of 89.27%, and other top-n results also significantly surpassed the baseline accuracy. Other experiment results demonstrate the superiority of the proposed modules and theirvaluable applications in ancient text intelligent recognition and cultural heritage digital preservation. Furthermore, this approach holds significant promise in facilitating the study of other handwritten or ancient characters recognition. Dataset and code are available at: <span><span>https://github.com/Nancy45-ui/WA-Net</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"140 ","pages":"Article 109674"},"PeriodicalIF":7.5,"publicationDate":"2024-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142700727","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}

引用次数: 0

Exploring structural components in autoencoder-based data clustering 探索基于自动编码器的数据聚类中的结构成分

IF 7.5 2区计算机科学 Q1 AUTOMATION & CONTROL SYSTEMS

Engineering Applications of Artificial Intelligence

Pub Date : 2024-11-25 DOI: 10.1016/j.engappai.2024.109562

Sujoy Chatterjee , Suvra Jyoti Choudhury

Clustering is a fundamental machine-learning task that has received extensive popularity in the literature. The foundational tenet of traditional clustering approaches is that data are learned to be vectorized features through various representational learning techniques. The conventional clustering methods can no longer manage the high-dimensional data as the data gets more intricate. Numerous representational learning strategies using deep architectures have been presented over the years, particularly deep unsupervised learning due to its superiority over conventional approaches. In most existing research, especially in the autoencoder-based approaches, only the distance information of pair-of-points in the original data space is retained in the latent space. However, combining this with additional preserving factors like the variance and independent component in the original data and latent space, respectively, is important. In addition, the model’s stability under noisy conditions is crucial. This paper provides a unique method for clustering data that combines autoencoder (AE), principal component analysis (PCA), and independent component analysis (ICA) to capture a relevant latent space representation. A further aid in lowering the dimensionality to improve clustering effectiveness is employing two-dimensional reduction algorithms, i.e., PCA and

t -

distributed stochastic neighbor embedding (

t -

SNE). The proposed technique produces more precise and reliable clustering by utilizing the advantages of both approaches. To compare the efficiency of the proposed methods to conventional clustering methods and stand-alone autoencoders, we conduct comprehensive experiments on 13 real-life datasets. The outcomes demonstrate the approach’s intriguing potential for addressing complicated clustering problems, and importantly, effectiveness is demonstrated even under noisy conditions.

聚类是一项基本的机器学习任务，在文献中广为流行。传统聚类方法的基本原理是通过各种表征学习技术将数据学习为向量化特征。随着数据变得越来越复杂，传统的聚类方法已无法管理高维数据。多年来，使用深度架构的表征学习策略层出不穷，尤其是深度无监督学习，因为它比传统方法更具优势。在大多数现有研究中，尤其是基于自动编码器的方法中，潜空间中只保留了原始数据空间中点对的距离信息。然而，将其分别与原始数据和潜空间中的方差和独立分量等额外的保留因子结合起来是非常重要的。此外，模型在噪声条件下的稳定性也至关重要。本文提供了一种独特的数据聚类方法，它结合了自动编码器（AE）、主成分分析（PCA）和独立成分分析（ICA）来捕捉相关的潜空间表示。为进一步降低维度以提高聚类效果，还采用了二维缩减算法，即 PCA 和 t 分布随机邻域嵌入（t-SNE）。所提出的技术利用了这两种方法的优势，能产生更精确、更可靠的聚类。为了比较所提方法与传统聚类方法和独立自动编码器的效率，我们在 13 个真实数据集上进行了全面实验。实验结果表明，该方法具有解决复杂聚类问题的巨大潜力，更重要的是，即使在噪声条件下，该方法的有效性也得到了证明。

{"title":"Exploring structural components in autoencoder-based data clustering","authors":"Sujoy Chatterjee , Suvra Jyoti Choudhury","doi":"10.1016/j.engappai.2024.109562","DOIUrl":"10.1016/j.engappai.2024.109562","url":null,"abstract":"<div><div>Clustering is a fundamental machine-learning task that has received extensive popularity in the literature. The foundational tenet of traditional clustering approaches is that data are learned to be vectorized features through various representational learning techniques. The conventional clustering methods can no longer manage the high-dimensional data as the data gets more intricate. Numerous representational learning strategies using deep architectures have been presented over the years, particularly deep unsupervised learning due to its superiority over conventional approaches. In most existing research, especially in the autoencoder-based approaches, only the distance information of pair-of-points in the original data space is retained in the latent space. However, combining this with additional preserving factors like the variance and independent component in the original data and latent space, respectively, is important. In addition, the model’s stability under noisy conditions is crucial. This paper provides a unique method for clustering data that combines autoencoder (AE), principal component analysis (PCA), and independent component analysis (ICA) to capture a relevant latent space representation. A further aid in lowering the dimensionality to improve clustering effectiveness is employing two-dimensional reduction algorithms, i.e., PCA and <span><math><mrow><mi>t</mi><mo>−</mo></mrow></math></span>distributed stochastic neighbor embedding (<span><math><mrow><mi>t</mi><mo>−</mo></mrow></math></span>SNE). The proposed technique produces more precise and reliable clustering by utilizing the advantages of both approaches. To compare the efficiency of the proposed methods to conventional clustering methods and stand-alone autoencoders, we conduct comprehensive experiments on 13 real-life datasets. The outcomes demonstrate the approach’s intriguing potential for addressing complicated clustering problems, and importantly, effectiveness is demonstrated even under noisy conditions.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"140 ","pages":"Article 109562"},"PeriodicalIF":7.5,"publicationDate":"2024-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142700796","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}

引用次数: 0

Motion planning for 7-degree-of-freedom bionic arm: Deep deterministic policy gradient algorithm based on imitation of human action 7 自由度仿生手臂的运动规划：基于模仿人类动作的深度确定性策略梯度算法

IF 7.5 2区计算机科学 Q1 AUTOMATION & CONTROL SYSTEMS

Engineering Applications of Artificial Intelligence

Pub Date : 2024-11-25 DOI: 10.1016/j.engappai.2024.109673

Baojiang Li , Shengjie Qiu , Haiyan Ye , Yuting Guo , Haiyan Wang , Jibo Bai

Smart bionic arms have played a great role in returning amputees to society. However, most of the current bionic arms are radial configuration type with few degrees of freedom and humeral form configuration type, which are only applicable to patients with proximal amputation. Patients with shoulder amputation urgently need a 7-degree-of-freedom bionic arm that can fully mimic human upper limb movements. Meanwhile, bionic arms often require specific programming to be implemented for the subject to initially meet the control requirements, which makes it difficult to match the motion of the bionic arm with the wearer's movement habits and reduces the adaptability and reliability of human-computer interaction. To address this problem, this paper proposes a motion imitation based on human upper limb joint point guidance and a motion planning algorithm based on reinforcement learning method to achieve the purpose of making the shoulder disconnected bionic arm accomplish humanoid motion by learning the dynamic motion imitation of the human upper limb. The algorithm analyzes and learns 3D poses of human arm movement features from unlabeled videos, then designs a reward function based on human motion patterns, and uses a reinforcement learning algorithm based on deep deterministic policy gradient (DDPG) to train the humanoid motion of the bionic arm. We evaluated the effectiveness of shoulder detached bionic arms through several tasks in a simulation environment, and the results showed that this method has good performance in planning the humanoid motion of bionic arms and can be widely applied in bionic machine control.

智能仿生臂在让截肢者重返社会方面发挥了巨大作用。然而，目前的仿生臂大多为自由度较小的桡骨构型和肱骨构型，仅适用于近端截肢的患者。肩部截肢患者迫切需要一种能完全模拟人类上肢运动的 7 自由度仿生手臂。同时，仿生手臂往往需要为受试者实施特定的编程才能初步满足控制要求，这使得仿生手臂的运动很难与佩戴者的运动习惯相匹配，降低了人机交互的适应性和可靠性。针对这一问题，本文提出了基于人体上肢关节点引导的运动模仿和基于强化学习方法的运动规划算法，通过学习人体上肢的动态运动模仿，达到使肩部断开的仿生手臂完成仿人运动的目的。该算法从未标明的视频中分析和学习人体手臂运动特征的三维姿势，然后根据人体运动模式设计奖励函数，并使用基于深度确定性策略梯度（DDPG）的强化学习算法来训练仿生手臂的仿人运动。我们在仿真环境中通过多个任务评估了肩部分离仿生手臂的有效性，结果表明该方法在规划仿生手臂的仿人运动方面具有良好的性能，可广泛应用于仿生机器控制领域。

{"title":"Motion planning for 7-degree-of-freedom bionic arm: Deep deterministic policy gradient algorithm based on imitation of human action","authors":"Baojiang Li , Shengjie Qiu , Haiyan Ye , Yuting Guo , Haiyan Wang , Jibo Bai","doi":"10.1016/j.engappai.2024.109673","DOIUrl":"10.1016/j.engappai.2024.109673","url":null,"abstract":"<div><div>Smart bionic arms have played a great role in returning amputees to society. However, most of the current bionic arms are radial configuration type with few degrees of freedom and humeral form configuration type, which are only applicable to patients with proximal amputation. Patients with shoulder amputation urgently need a 7-degree-of-freedom bionic arm that can fully mimic human upper limb movements. Meanwhile, bionic arms often require specific programming to be implemented for the subject to initially meet the control requirements, which makes it difficult to match the motion of the bionic arm with the wearer's movement habits and reduces the adaptability and reliability of human-computer interaction. To address this problem, this paper proposes a motion imitation based on human upper limb joint point guidance and a motion planning algorithm based on reinforcement learning method to achieve the purpose of making the shoulder disconnected bionic arm accomplish humanoid motion by learning the dynamic motion imitation of the human upper limb. The algorithm analyzes and learns 3D poses of human arm movement features from unlabeled videos, then designs a reward function based on human motion patterns, and uses a reinforcement learning algorithm based on deep deterministic policy gradient (DDPG) to train the humanoid motion of the bionic arm. We evaluated the effectiveness of shoulder detached bionic arms through several tasks in a simulation environment, and the results showed that this method has good performance in planning the humanoid motion of bionic arms and can be widely applied in bionic machine control.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"140 ","pages":"Article 109673"},"PeriodicalIF":7.5,"publicationDate":"2024-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142700720","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}

引用次数: 0

A return-to-home unmanned aerial vehicle navigation solution in global positioning system denied environments via bidirectional long short-term memory reverse flightpath prediction 通过双向长短期记忆反向飞行路径预测，在全球定位系统拒绝的环境中实现无人飞行器返航导航解决方案

IF 7.5 2区计算机科学 Q1 AUTOMATION & CONTROL SYSTEMS

Engineering Applications of Artificial Intelligence

Pub Date : 2024-11-25 DOI: 10.1016/j.engappai.2024.109729

Mustafa Alkhatib , Mohammad Nayfeh , Khair Al Shamaileh , Naima Kaabouch , Vijay Devabhaktuni

In this paper, bidirectional long short-term memory (B-LSTM) deep learning modeling is proposed as an approach to facilitate autonomous return-to-home (RTH) aerial navigation in environments with compromised global positioning system (GPS) reception. Logged samples of ten radiometric features are extracted from onboard sensors (i.e., accelerometer, barometer, GPS, gyroscope, magnetometer) in two outdoor experimental scenarios of different altitudes and velocities. These samples are used for training and validating B-LSTM models with single and parallel architectures. The former architecture consists of a single B-LSTM model that processes all input features across the x-, y-, and z-axes to predict a three-dimensional local position, whereas the latter comprises three parallel B-LSTM models, each for processing only the features of a specific dimension (i.e., x, y, or z) and predicting local position in the respective axis. Evaluations demonstrate the validity of the proposed approach, with a 4-m average mean square error (MSE). This is achieved without imposing resource-consuming computational overhead, modifications to existing hardware, or changes to physical infrastructure and communication protocols. Due to using existing onboard sensors and accommodating varied scenarios, the proposed approach finds applications in autonomous navigation, including unmanned aerial vehicles (UAVs) and ground vehicles.

本文提出了一种双向长短期记忆（B-LSTM）深度学习建模方法，用于在全球定位系统（GPS）接收受到影响的环境中促进自主返回原点（RTH）空中导航。在两个不同高度和速度的室外实验场景中，从机载传感器（即加速度计、气压计、全球定位系统、陀螺仪、磁力计）中提取了十个辐射特征的记录样本。这些样本用于训练和验证采用单一和并行架构的 B-LSTM 模型。前一种架构由一个 B-LSTM 模型组成，该模型处理 x、y 和 z 轴上的所有输入特征，以预测三维局部位置；而后一种架构由三个并行 B-LSTM 模型组成，每个模型只处理特定维度（即 x、y 或 z 轴）的特征，并预测相应轴上的局部位置。评估证明了所提方法的有效性，平均均方误差 (MSE) 为 4 米。实现这一目标不需要耗费资源的计算开销，不需要修改现有硬件，也不需要改变物理基础设施和通信协议。由于使用了现有的机载传感器并能适应不同的场景，所提出的方法可应用于自主导航，包括无人驾驶飞行器（UAV）和地面车辆。

{"title":"A return-to-home unmanned aerial vehicle navigation solution in global positioning system denied environments via bidirectional long short-term memory reverse flightpath prediction","authors":"Mustafa Alkhatib , Mohammad Nayfeh , Khair Al Shamaileh , Naima Kaabouch , Vijay Devabhaktuni","doi":"10.1016/j.engappai.2024.109729","DOIUrl":"10.1016/j.engappai.2024.109729","url":null,"abstract":"<div><div>In this paper, bidirectional long short-term memory (B-LSTM) deep learning modeling is proposed as an approach to facilitate autonomous return-to-home (RTH) aerial navigation in environments with compromised global positioning system (GPS) reception. Logged samples of ten radiometric features are extracted from onboard sensors (i.e., accelerometer, barometer, GPS, gyroscope, magnetometer) in two outdoor experimental scenarios of different altitudes and velocities. These samples are used for training and validating B-LSTM models with single and parallel architectures. The former architecture consists of a single B-LSTM model that processes all input features across the <em>x</em>-, <em>y</em>-, and <em>z</em>-axes to predict a three-dimensional local position, whereas the latter comprises three parallel B-LSTM models, each for processing only the features of a specific dimension (i.e., <em>x</em>, <em>y</em>, or <em>z</em>) and predicting local position in the respective axis. Evaluations demonstrate the validity of the proposed approach, with a 4-m average mean square error (MSE). This is achieved without imposing resource-consuming computational overhead, modifications to existing hardware, or changes to physical infrastructure and communication protocols. Due to using existing onboard sensors and accommodating varied scenarios, the proposed approach finds applications in autonomous navigation, including unmanned aerial vehicles (UAVs) and ground vehicles.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"140 ","pages":"Article 109729"},"PeriodicalIF":7.5,"publicationDate":"2024-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142700772","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}

引用次数: 0

A depthwise convolutional neural network model based on active contour for multi-defect wafer map pattern classification 基于主动轮廓的深度卷积神经网络模型，用于多缺陷晶片图模式分类

IF 7.5 2区计算机科学 Q1 AUTOMATION & CONTROL SYSTEMS

Engineering Applications of Artificial Intelligence

Pub Date : 2024-11-24 DOI: 10.1016/j.engappai.2024.109707

Jeonghoon Choi, Dongjun Suh

As semiconductor manufacturing processes continue to witness increased integration density and design complexity, semiconductor wafers are experiencing a growing diversity and complexity of defects. While previous research in wafer map classification using deep learning has made significant advancements in dealing with single defect patterns, the classification of mixed-type defects has received less attention due to their considerably higher difficulty level compared to single defects. This research addresses this critical gap, emphasizing the need for improved methods to classify mixed-type defects, which are more complex and challenging. To tackle this challenge, this paper introduces the active contour-based lightweight depthwise network (AC-LDN) model for the classification of multi-defect wafer map patterns. Initially, multi-defect features are extracted using an active contour-based segmentation model. Subsequently, the learning model employs a depthwise convolutional neural network (CNN) architecture that combines separable CNN and dilated CNN techniques. This unique approach optimizes the model in the separable segment while effectively addressing defect complexity in the depthwise segments. Consequently, AC-LDN outperforms other state-of-the-art models, offering a balance between lightweight characteristics and high accuracy. The proposed method demonstrates its superiority over previous models when evaluated on the extsdsensive multi-wafer map dataset, achieving an average classification accuracy exceeding 98% and a confusion matrix coefficient surpassing 0.97.

随着半导体制造工艺的集成密度和设计复杂性不断提高，半导体晶片缺陷的多样性和复杂性也在不断增加。以往利用深度学习进行晶圆图分类的研究在处理单一缺陷模式方面取得了重大进展，但混合型缺陷的分类却较少受到关注，因为与单一缺陷相比，混合型缺陷的难度要高得多。本研究弥补了这一重大缺陷，强调需要改进混合型缺陷的分类方法，因为混合型缺陷更为复杂，更具挑战性。为了应对这一挑战，本文引入了基于主动轮廓的轻量级深度网络（AC-LDN）模型，用于多缺陷晶片图模式的分类。首先，使用基于主动轮廓的分割模型提取多缺陷特征。随后，学习模型采用深度卷积神经网络（CNN）架构，该架构结合了可分离 CNN 和扩张 CNN 技术。这种独特的方法优化了可分离分段中的模型，同时有效解决了深度分段中的缺陷复杂性。因此，AC-LDN 超越了其他最先进的模型，在轻量级特性和高精度之间取得了平衡。在外部密集型多晶片地图数据集上进行评估时，所提出的方法证明了其优于以往模型的性能，平均分类准确率超过 98%，混淆矩阵系数超过 0.97。

{"title":"A depthwise convolutional neural network model based on active contour for multi-defect wafer map pattern classification","authors":"Jeonghoon Choi, Dongjun Suh","doi":"10.1016/j.engappai.2024.109707","DOIUrl":"10.1016/j.engappai.2024.109707","url":null,"abstract":"<div><div>As semiconductor manufacturing processes continue to witness increased integration density and design complexity, semiconductor wafers are experiencing a growing diversity and complexity of defects. While previous research in wafer map classification using deep learning has made significant advancements in dealing with single defect patterns, the classification of mixed-type defects has received less attention due to their considerably higher difficulty level compared to single defects. This research addresses this critical gap, emphasizing the need for improved methods to classify mixed-type defects, which are more complex and challenging. To tackle this challenge, this paper introduces the active contour-based lightweight depthwise network (AC-LDN) model for the classification of multi-defect wafer map patterns. Initially, multi-defect features are extracted using an active contour-based segmentation model. Subsequently, the learning model employs a depthwise convolutional neural network (CNN) architecture that combines separable CNN and dilated CNN techniques. This unique approach optimizes the model in the separable segment while effectively addressing defect complexity in the depthwise segments. Consequently, AC-LDN outperforms other state-of-the-art models, offering a balance between lightweight characteristics and high accuracy. The proposed method demonstrates its superiority over previous models when evaluated on the extsdsensive multi-wafer map dataset, achieving an average classification accuracy exceeding 98% and a confusion matrix coefficient surpassing 0.97.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"139 ","pages":"Article 109707"},"PeriodicalIF":7.5,"publicationDate":"2024-11-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142720677","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}

引用次数: 0