Pub Date : 2025-02-04DOI: 10.1109/THMS.2025.3526267
{"title":"Call for Papers: IEEE Transactions on Human-Machine Systems","authors":"","doi":"10.1109/THMS.2025.3526267","DOIUrl":"https://doi.org/10.1109/THMS.2025.3526267","url":null,"abstract":"","PeriodicalId":48916,"journal":{"name":"IEEE Transactions on Human-Machine Systems","volume":"55 1","pages":"112-112"},"PeriodicalIF":3.5,"publicationDate":"2025-02-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10871203","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143106816","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-04DOI: 10.1109/THMS.2024.3523663
{"title":"IEEE Transactions on Human-Machine Systems Information for Authors","authors":"","doi":"10.1109/THMS.2024.3523663","DOIUrl":"https://doi.org/10.1109/THMS.2024.3523663","url":null,"abstract":"","PeriodicalId":48916,"journal":{"name":"IEEE Transactions on Human-Machine Systems","volume":"55 1","pages":"C4-C4"},"PeriodicalIF":3.5,"publicationDate":"2025-02-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10871229","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143106817","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-04DOI: 10.1109/THMS.2024.3523659
{"title":"IEEE Systems, Man, and Cybernetics Society Information","authors":"","doi":"10.1109/THMS.2024.3523659","DOIUrl":"https://doi.org/10.1109/THMS.2024.3523659","url":null,"abstract":"","PeriodicalId":48916,"journal":{"name":"IEEE Transactions on Human-Machine Systems","volume":"55 1","pages":"C2-C2"},"PeriodicalIF":3.5,"publicationDate":"2025-02-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10871175","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143106857","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-03DOI: 10.1109/THMS.2025.3527136
Leo Julius Materne;Maik Friedrich
Remote air traffic control offers inexpensive and efficient service to multiple airports. Recent research shows that one remote air traffic control officer can safely control up to three low-traffic airports simultaneously. In a multiple remote tower center, airports can be flexibly allocated across air traffic control officers based on prospective traffic loads. The main task of the supervisor in such a center is balancing the workload of each air traffic control officer by allocating airports accordingly. This study analyzes the supervisor's visual attention during interaction with a planning tool for their daily tasks. Five use cases were identified as the main tasks of the supervisor representing a mixture of planned and unplanned events. A mixed methods within-subjects design was used to assess the workload and eye-movement patterns associated with each of these use cases. In total, 15 professional air traffic control officers participated in the study. Workload and eye movement were analyzed independently in relation to the use cases but also in combination with each other. Across all use cases, a small correlation between subjective workload ratings and fixation duration was found, supporting previous findings of fixation duration being associated with information processing. Transitions between areas of interest on the supervisor planning tool provided valuable insights into the layout design of future supervisor planning tools.
{"title":"Supervision of Multiple Remote Tower Centers: Evaluating a New Air Traffic Control Interface Based on Mental Workload and Eye Tracking","authors":"Leo Julius Materne;Maik Friedrich","doi":"10.1109/THMS.2025.3527136","DOIUrl":"https://doi.org/10.1109/THMS.2025.3527136","url":null,"abstract":"Remote air traffic control offers inexpensive and efficient service to multiple airports. Recent research shows that one remote air traffic control officer can safely control up to three low-traffic airports simultaneously. In a multiple remote tower center, airports can be flexibly allocated across air traffic control officers based on prospective traffic loads. The main task of the supervisor in such a center is balancing the workload of each air traffic control officer by allocating airports accordingly. This study analyzes the supervisor's visual attention during interaction with a planning tool for their daily tasks. Five use cases were identified as the main tasks of the supervisor representing a mixture of planned and unplanned events. A mixed methods within-subjects design was used to assess the workload and eye-movement patterns associated with each of these use cases. In total, 15 professional air traffic control officers participated in the study. Workload and eye movement were analyzed independently in relation to the use cases but also in combination with each other. Across all use cases, a small correlation between subjective workload ratings and fixation duration was found, supporting previous findings of fixation duration being associated with information processing. Transitions between areas of interest on the supervisor planning tool provided valuable insights into the layout design of future supervisor planning tools.","PeriodicalId":48916,"journal":{"name":"IEEE Transactions on Human-Machine Systems","volume":"55 2","pages":"114-123"},"PeriodicalIF":3.5,"publicationDate":"2025-02-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10869643","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143698311","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-03DOI: 10.1109/THMS.2025.3529759
Mohammad Soleimani Amiri;Rizauddin Ramli
The motivation behind the development of a wearable assistive upper limb exoskeleton robot was to provide comprehensive multijoint therapy by assisting physiotherapists in enhancing the recovery of hemiplegic patients. However, the controlling of an upper limb exoskeleton for rehabilitation is a challenging task because of its nonlinear characteristics. This article presents a novel fuzzy adaptive controller that utilizes a high-dimensional integral-type Lyapunov function for a wearable assistive upper limb exoskeleton. A disturbance observer had been used to tackle uncertainties in the exoskeleton's dynamic model, thereby enhancing the tracking performance of the joints. The aim of this control scheme was to overcome unknown parameters in the dynamic model. The performance of the adaptive controller was validated through human interactive experiments and periodically repeated reference trajectory tests. The results demonstrated that the proposed fuzzy adaptive controller, with the inclusion of a disturbance observer, could effectively compensate for uncertain disturbances and could achieve efficient tracking of the reference trajectory. The statistical analysis revealed that the fuzzy adaptive controller performed 45%, 44%, and 31% less in average error compared to adaptive conventional controllers. The findings ascertained the potential of the proposed controller in improving the recovery of motor functions of hemiplegic patients.
{"title":"Fuzzy Adaptive Controller of a Wearable Assistive Upper Limb Exoskeleton Using a Disturbance Observer","authors":"Mohammad Soleimani Amiri;Rizauddin Ramli","doi":"10.1109/THMS.2025.3529759","DOIUrl":"https://doi.org/10.1109/THMS.2025.3529759","url":null,"abstract":"The motivation behind the development of a wearable assistive upper limb exoskeleton robot was to provide comprehensive multijoint therapy by assisting physiotherapists in enhancing the recovery of hemiplegic patients. However, the controlling of an upper limb exoskeleton for rehabilitation is a challenging task because of its nonlinear characteristics. This article presents a novel fuzzy adaptive controller that utilizes a high-dimensional integral-type Lyapunov function for a wearable assistive upper limb exoskeleton. A disturbance observer had been used to tackle uncertainties in the exoskeleton's dynamic model, thereby enhancing the tracking performance of the joints. The aim of this control scheme was to overcome unknown parameters in the dynamic model. The performance of the adaptive controller was validated through human interactive experiments and periodically repeated reference trajectory tests. The results demonstrated that the proposed fuzzy adaptive controller, with the inclusion of a disturbance observer, could effectively compensate for uncertain disturbances and could achieve efficient tracking of the reference trajectory. The statistical analysis revealed that the fuzzy adaptive controller performed 45%, 44%, and 31% less in average error compared to adaptive conventional controllers. The findings ascertained the potential of the proposed controller in improving the recovery of motor functions of hemiplegic patients.","PeriodicalId":48916,"journal":{"name":"IEEE Transactions on Human-Machine Systems","volume":"55 2","pages":"197-206"},"PeriodicalIF":3.5,"publicationDate":"2025-02-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143698219","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This article introduces the global-local image perceptual score (GLIPS), an image metric designed to assess the photorealistic image quality of AI-generated images with a high degree of alignment to human visual perception. Traditional metrics such as Fréchet inception distance (FID) and kernel inception distance scores do not align closely with human evaluations. The proposed metric incorporates advanced transformer-based attention mechanisms to assess local similarity and maximum mean discrepancy to evaluate global distributional similarity. To evaluate the performance of GLIPS, we conducted a human study on photorealistic image quality. Comprehensive tests across various generative models demonstrate that GLIPS consistently outperforms existing metrics like FID, structural similarity index measure, and multiscale structural similarity index measure in terms of correlation with human scores. In addition, we introduce the interpolative binning scale, a refined scaling method that enhances the interpretability of metric scores by aligning them more closely with human evaluative standards. The proposed metric and scaling approach not only provide more reliable assessments of AI-generated images but also suggest pathways for future enhancements in image generation technologies.
{"title":"Global-Local Image Perceptual Score (GLIPS): Evaluating Photorealistic Quality of AI-Generated Images","authors":"Memoona Aziz;Umair Rehman;Muhammad Umair Danish;Katarina Grolinger","doi":"10.1109/THMS.2025.3527397","DOIUrl":"https://doi.org/10.1109/THMS.2025.3527397","url":null,"abstract":"This article introduces the global-local image perceptual score (GLIPS), an image metric designed to assess the photorealistic image quality of AI-generated images with a high degree of alignment to human visual perception. Traditional metrics such as Fréchet inception distance (FID) and kernel inception distance scores do not align closely with human evaluations. The proposed metric incorporates advanced transformer-based attention mechanisms to assess local similarity and maximum mean discrepancy to evaluate global distributional similarity. To evaluate the performance of GLIPS, we conducted a human study on photorealistic image quality. Comprehensive tests across various generative models demonstrate that GLIPS consistently outperforms existing metrics like FID, structural similarity index measure, and multiscale structural similarity index measure in terms of correlation with human scores. In addition, we introduce the interpolative binning scale, a refined scaling method that enhances the interpretability of metric scores by aligning them more closely with human evaluative standards. The proposed metric and scaling approach not only provide more reliable assessments of AI-generated images but also suggest pathways for future enhancements in image generation technologies.","PeriodicalId":48916,"journal":{"name":"IEEE Transactions on Human-Machine Systems","volume":"55 2","pages":"223-233"},"PeriodicalIF":3.5,"publicationDate":"2025-02-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143698312","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-29DOI: 10.1109/THMS.2025.3526957
Ahmad Chaddad
Radiomics combined with deep learning is an emerging field within biomedical engineering that aims to extract important characteristics from medical images to develop a predictive model that can support clinical decision-making. This method could be used in the realm of brain disorders, particularly autism spectrum disorder (ASD), to facilitate prompt identification. We propose a novel radiomic features [deep radiomic features (DTF)], involving the use of principal component analysis to encode convolutional neural network (CNN) features, thereby capturing distinctive features related to brain regions in subjects with ASD subjects and their age. Using these features in random forest (RF) models, we explore two scenarios, such as site-specific radiomic analysis and feature extraction from unaffected brain regions to alleviate site-related variations. Our experiments involved comparing the proposed method with standard radiomics (SR) and 2-D/3-D CNNs for the classification of ASD versus healthy control (HC) individuals and different age groups (below median and above median). When using the RF model with DTF, the analysis at individual sites revealed an area under the receiver operating characteristic (ROC) curve (AUC) range of 79%–85% for features, such as the left lateral-ventricle, cerebellum-white-matter, and pallidum, as well as the right choroid-plexus and vessel. In the context of fivefold cross validation with the RF model, the combined features (DTF from 3-D CNN, ResNet50, DarketNet53, and NasNet_large with SR) achieved the highest AUC value of 76.67%. Furthermore, our method also showed notable AUC values for predicting age in subjects with ASD (80.91%) and HC (75.64%). The results indicate that DTFs consistently exhibit predictive value in classifying ASD from HC subjects and in predicting age.
放射组学与深度学习相结合是生物医学工程中的一个新兴领域，旨在从医学图像中提取重要特征，以开发可支持临床决策的预测模型。该方法可用于脑部疾病领域，特别是自闭症谱系障碍（ASD），以促进及时识别。我们提出了一种新的放射学特征[deep radiomic features (DTF)]，涉及使用主成分分析来编码卷积神经网络（CNN）特征，从而捕获与ASD受试者及其年龄相关的大脑区域的独特特征。利用随机森林（RF）模型中的这些特征，我们探索了两种场景，即特定位点的放射学分析和未受影响的大脑区域的特征提取，以减轻位点相关的变异。我们的实验涉及将所提出的方法与标准放射组学（SR）和2-D/3-D cnn进行比较，以区分ASD与健康对照（HC）个体和不同年龄组（低于中位数和高于中位数）。当使用RF模型和DTF时，单个部位的分析显示，左侧侧脑室、小脑-白质、白质以及右侧脉络丛和血管等特征的接受者工作特征曲线（ROC）下面积（AUC）范围为79%-85%。在与RF模型进行五重交叉验证的情况下，组合特征（来自3-D CNN、ResNet50、DarketNet53和NasNet_large的DTF与SR）的AUC值最高，为76.67%。此外，我们的方法在预测ASD（80.91%）和HC（75.64%）患者的年龄方面也显示出显著的AUC值。结果表明，DTFs对HC患者的ASD分类和年龄预测具有一致的预测价值。
{"title":"Deep Radiomics for Autism Diagnosis and Age Prediction","authors":"Ahmad Chaddad","doi":"10.1109/THMS.2025.3526957","DOIUrl":"https://doi.org/10.1109/THMS.2025.3526957","url":null,"abstract":"Radiomics combined with deep learning is an emerging field within biomedical engineering that aims to extract important characteristics from medical images to develop a predictive model that can support clinical decision-making. This method could be used in the realm of brain disorders, particularly autism spectrum disorder (ASD), to facilitate prompt identification. We propose a novel radiomic features [deep radiomic features (DTF)], involving the use of principal component analysis to encode convolutional neural network (CNN) features, thereby capturing distinctive features related to brain regions in subjects with ASD subjects and their age. Using these features in random forest (RF) models, we explore two scenarios, such as site-specific radiomic analysis and feature extraction from unaffected brain regions to alleviate site-related variations. Our experiments involved comparing the proposed method with standard radiomics (SR) and 2-D/3-D CNNs for the classification of ASD versus healthy control (HC) individuals and different age groups (below median and above median). When using the RF model with DTF, the analysis at individual sites revealed an area under the receiver operating characteristic (ROC) curve (AUC) range of 79%–85% for features, such as the left <italic>lateral-ventricle</i>, <italic>cerebellum-white-matter,</i> and <italic>pallidum</i>, as well as the right <italic>choroid-plexus</i> and <italic>vessel</i>. In the context of fivefold cross validation with the RF model, the combined features (DTF from 3-D CNN, ResNet50, DarketNet53, and NasNet_large with SR) achieved the highest AUC value of 76.67%. Furthermore, our method also showed notable AUC values for predicting age in subjects with ASD (80.91%) and HC (75.64%). The results indicate that DTFs consistently exhibit predictive value in classifying ASD from HC subjects and in predicting age.","PeriodicalId":48916,"journal":{"name":"IEEE Transactions on Human-Machine Systems","volume":"55 2","pages":"144-154"},"PeriodicalIF":3.5,"publicationDate":"2025-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143698220","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-22DOI: 10.1109/THMS.2024.3524916
Kym K. W. Man;Jeremy A. Patterson;Christopher T. Simons
In sensory science, the use of immersive technologies has gained popularity for their ability to restore relevant contextual factors during consumer testing and overcome the low ecological validity of controlled laboratory environments. Despite this, there is scant literature evaluating the effectiveness of immersive technologies in facilitating virtual product evaluation experiences; this is especially true with virtual reality (VR) headsets and the unique technical challenges associated with this technology. To fill this gap, we assessed virtual presence, system usability, engagement, and ease of task completion, in subjects using two iterations of a VR application (controllers or hand tracking) designed to address the major limitations of current systems. Results revealed that both systems exceeded the benchmark usability score of 68. System 1 (controllers) performed better for interactions with the virtual tablet interface to answer questions, whereas interactions with the food objects were easier using System 2 (hand tracking). Participants also experienced a high sense of virtual presence using both systems. When measured in System 2, a high level of subject engagement during the immersive product evaluations was observed. These studies indicate that collecting both quantitative and qualitative feedback on VR systems can provide useful insights and directions for application optimization to ensure valid investigation of context effects in future research.
{"title":"Efficacy Assessments of Virtual Reality Systems for Immersive Consumer Testing—Two Case Studies With Tortilla Chip Evaluations","authors":"Kym K. W. Man;Jeremy A. Patterson;Christopher T. Simons","doi":"10.1109/THMS.2024.3524916","DOIUrl":"https://doi.org/10.1109/THMS.2024.3524916","url":null,"abstract":"In sensory science, the use of immersive technologies has gained popularity for their ability to restore relevant contextual factors during consumer testing and overcome the low ecological validity of controlled laboratory environments. Despite this, there is scant literature evaluating the effectiveness of immersive technologies in facilitating virtual product evaluation experiences; this is especially true with virtual reality (VR) headsets and the unique technical challenges associated with this technology. To fill this gap, we assessed virtual presence, system usability, engagement, and ease of task completion, in subjects using two iterations of a VR application (controllers or hand tracking) designed to address the major limitations of current systems. Results revealed that both systems exceeded the benchmark usability score of 68. System 1 (controllers) performed better for interactions with the virtual tablet interface to answer questions, whereas interactions with the food objects were easier using System 2 (hand tracking). Participants also experienced a high sense of virtual presence using both systems. When measured in System 2, a high level of subject engagement during the immersive product evaluations was observed. These studies indicate that collecting both quantitative and qualitative feedback on VR systems can provide useful insights and directions for application optimization to ensure valid investigation of context effects in future research.","PeriodicalId":48916,"journal":{"name":"IEEE Transactions on Human-Machine Systems","volume":"55 2","pages":"266-277"},"PeriodicalIF":3.5,"publicationDate":"2025-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143698299","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-17DOI: 10.1109/THMS.2024.3522974
Mengyuan Liu;Chen Chen;Songtao Wu;Fanyang Meng;Hong Liu
Recognizing interactive actions, including hand-to-hand interaction and human-to-human interaction, has attracted increasing attention for various applications in the field of video analysis and human–robot interaction. Considering the success of graph convolution in modeling topology-aware features from skeleton data, recent methods commonly operate graph convolution on separate entities and use late fusion for interactive action recognition, which can barely model the mutual semantic relationships between pairwise entities. To this end, we propose a mutual excitation graph convolutional network (me-GCN) by stacking mutual excitation graph convolution (me-GC) layers. Specifically, me-GC uses a mutual topology excitation module to firstly extract adjacency matrices from individual entities and then adaptively model the mutual constraints between them. Moreover, me-GC extends the above idea and further uses a mutual feature excitation module to extract and merge deep features from pairwise entities. Compared with graph convolution, our proposed me-GC gradually learns mutual information in each layer and each stage of graph convolution operations. Extensive experiments on a challenging hand-to-hand interaction dataset, i.e., the Assembely101 dataset, and two large-scale human-to-human interaction datasets, i.e., NTU60-Interaction and NTU120-Interaction consistently verify the superiority of our proposed method, which outperforms the state-of-the-art GCN-based and Transformer-based methods.
{"title":"Learning Mutual Excitation for Hand-to-Hand and Human-to-Human Interaction Recognition","authors":"Mengyuan Liu;Chen Chen;Songtao Wu;Fanyang Meng;Hong Liu","doi":"10.1109/THMS.2024.3522974","DOIUrl":"https://doi.org/10.1109/THMS.2024.3522974","url":null,"abstract":"Recognizing interactive actions, including hand-to-hand interaction and human-to-human interaction, has attracted increasing attention for various applications in the field of video analysis and human–robot interaction. Considering the success of graph convolution in modeling topology-aware features from skeleton data, recent methods commonly operate graph convolution on separate entities and use late fusion for interactive action recognition, which can barely model the mutual semantic relationships between pairwise entities. To this end, we propose a mutual excitation graph convolutional network (me-GCN) by stacking mutual excitation graph convolution (me-GC) layers. Specifically, me-GC uses a mutual topology excitation module to firstly extract adjacency matrices from individual entities and then adaptively model the mutual constraints between them. Moreover, me-GC extends the above idea and further uses a mutual feature excitation module to extract and merge deep features from pairwise entities. Compared with graph convolution, our proposed me-GC gradually learns mutual information in each layer and each stage of graph convolution operations. Extensive experiments on a challenging hand-to-hand interaction dataset, i.e., the Assembely101 dataset, and two large-scale human-to-human interaction datasets, i.e., NTU60-Interaction and NTU120-Interaction consistently verify the superiority of our proposed method, which outperforms the state-of-the-art GCN-based and Transformer-based methods.","PeriodicalId":48916,"journal":{"name":"IEEE Transactions on Human-Machine Systems","volume":"55 2","pages":"134-143"},"PeriodicalIF":3.5,"publicationDate":"2025-01-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143698216","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-16DOI: 10.1109/THMS.2024.3496899
Xinxin Yao;Jiahang Liu;Xinglong Zhang;Xin Xu
Human–machine shared control has recently been regarded as a promising paradigm to improve safety and performance in complex driving scenarios. One crucial task in shared control is dynamically optimizing the driving weights between the driver and the intelligent vehicle to adapt to dynamic driving scenarios. However, designing an optimal human–machine shared controller with guaranteed performance and stability is challenging due to nonnegligible time delays caused by communication protocols and uncertainties in driver behavior. This article proposes a novel receding-horizon reinforcement learning approach for time-delayed human–machine shared control of intelligent vehicles. First, we build a multikernel-based data-driven model of vehicle dynamics and driving behavior, considering time delays and uncertainties of drivers' actions. Second, a model-based receding horizon actor–critic learning algorithm is presented to learn an explicit policy for time-delayed human–machine shared control online. Unlike classic reinforcement learning, policy learning of the proposed approach is performed according to a receding-horizon strategy to enhance learning efficiency and adaptability. In theory, the closed-loop stability under time delays is analyzed. Hardware-in-the-loop experiments on the time-delayed human–machine shared control of intelligent vehicles have been conducted in variable curvature road scenarios. The results demonstrate that our approach has significant improvements in driving performance and driver workload compared with pure manual driving and previous shared control methods.
{"title":"Receding-Horizon Reinforcement Learning for Time-Delayed Human–Machine Shared Control of Intelligent Vehicles","authors":"Xinxin Yao;Jiahang Liu;Xinglong Zhang;Xin Xu","doi":"10.1109/THMS.2024.3496899","DOIUrl":"https://doi.org/10.1109/THMS.2024.3496899","url":null,"abstract":"Human–machine shared control has recently been regarded as a promising paradigm to improve safety and performance in complex driving scenarios. One crucial task in shared control is dynamically optimizing the driving weights between the driver and the intelligent vehicle to adapt to dynamic driving scenarios. However, designing an optimal human–machine shared controller with guaranteed performance and stability is challenging due to nonnegligible time delays caused by communication protocols and uncertainties in driver behavior. This article proposes a novel receding-horizon reinforcement learning approach for time-delayed human–machine shared control of intelligent vehicles. First, we build a multikernel-based data-driven model of vehicle dynamics and driving behavior, considering time delays and uncertainties of drivers' actions. Second, a model-based receding horizon actor–critic learning algorithm is presented to learn an explicit policy for time-delayed human–machine shared control online. Unlike classic reinforcement learning, policy learning of the proposed approach is performed according to a receding-horizon strategy to enhance learning efficiency and adaptability. In theory, the closed-loop stability under time delays is analyzed. Hardware-in-the-loop experiments on the time-delayed human–machine shared control of intelligent vehicles have been conducted in variable curvature road scenarios. The results demonstrate that our approach has significant improvements in driving performance and driver workload compared with pure manual driving and previous shared control methods.","PeriodicalId":48916,"journal":{"name":"IEEE Transactions on Human-Machine Systems","volume":"55 2","pages":"155-164"},"PeriodicalIF":3.5,"publicationDate":"2025-01-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143698298","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：