The tactile object recognition (TOR) is highly important for environmental perception of robots. The previous works usually utilize single scale convolution which cannot simultaneously extract local and global spatiotemporal features of tactile data, which leads to low accuracy in TOR task. To address above problem, this article proposes a local and global residual (LGR-18) network which is mainly consisted of multiple local and global convolution (LGC) blocks. An LGC block contains two pairs of local convolution (LC) and global convolution (GC) modules. The LC module mainly utilizes a temporal shift operation and a 2D convolution layer to extract local spatiotemporal features. The GC module extracts global spatiotemporal features by fusing multiple 1D and 2D convolutions which can expand the receptive field in temporal and spatial dimensions. Consequently, our LGR-18 network can extract local-global spatiotemporal features without using 3D convolutions which usually require a large number of parameters. The effectiveness of LC module, GC module and LGC block is verified by ablation studies. Quantitative comparisons with state-of-the-art methods reveal the excellent capability of our method.
{"title":"Mining local and global spatiotemporal features for tactile object recognition","authors":"Xiaoliang Qian, Wei Deng, Wei Wang, Yucui Liu, Liying Jiang","doi":"10.3389/fnbot.2024.1387428","DOIUrl":"https://doi.org/10.3389/fnbot.2024.1387428","url":null,"abstract":"The tactile object recognition (TOR) is highly important for environmental perception of robots. The previous works usually utilize single scale convolution which cannot simultaneously extract local and global spatiotemporal features of tactile data, which leads to low accuracy in TOR task. To address above problem, this article proposes a local and global residual (LGR-18) network which is mainly consisted of multiple local and global convolution (LGC) blocks. An LGC block contains two pairs of local convolution (LC) and global convolution (GC) modules. The LC module mainly utilizes a temporal shift operation and a 2D convolution layer to extract local spatiotemporal features. The GC module extracts global spatiotemporal features by fusing multiple 1D and 2D convolutions which can expand the receptive field in temporal and spatial dimensions. Consequently, our LGR-18 network can extract local-global spatiotemporal features without using 3D convolutions which usually require a large number of parameters. The effectiveness of LC module, GC module and LGC block is verified by ablation studies. Quantitative comparisons with state-of-the-art methods reveal the excellent capability of our method.","PeriodicalId":12628,"journal":{"name":"Frontiers in Neurorobotics","volume":null,"pages":null},"PeriodicalIF":3.1,"publicationDate":"2024-05-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140841107","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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 : 2024-05-03DOI: 10.3389/fnbot.2024.1395652
Shaoqi Yu, Yintong Wang, Lili Chen, Xiaolin Zhang, Jiamao Li
In Human-Robot Interaction (HRI), accurate 3D hand pose and mesh estimation hold critical importance. However, inferring reasonable and accurate poses in severe self-occlusion and high self-similarity remains an inherent challenge. In order to alleviate the ambiguity caused by invisible and similar joints during HRI, we propose a new Topology-aware Transformer network named HandGCNFormer with depth image as input, incorporating prior knowledge of hand kinematic topology into the network while modeling long-range contextual information. Specifically, we propose a novel Graphformer decoder with an additional Node-offset Graph Convolutional layer (NoffGConv). The Graphformer decoder optimizes the synergy between the Transformer and GCN, capturing long-range dependencies and local topological connections between joints. On top of that, we replace the standard MLP prediction head with a novel Topology-aware head to better exploit local topological constraints for more reasonable and accurate poses. Our method achieves state-of-the-art 3D hand pose estimation performance on four challenging datasets, including Hands2017, NYU, ICVL, and MSRA. To further demonstrate the effectiveness and scalability of our proposed Graphformer Decoder and Topology aware head, we extend our framework to HandGCNFormer-Mesh for the 3D hand mesh estimation task. The extended framework efficiently integrates a shape regressor with the original Graphformer Decoder and Topology aware head, producing Mano parameters. The results on the HO-3D dataset, which contains various and challenging occlusions, show that our HandGCNFormer-Mesh achieves competitive results compared to previous state-of-the-art 3D hand mesh estimation methods.
{"title":"3D hand pose and mesh estimation via a generic Topology-aware Transformer model","authors":"Shaoqi Yu, Yintong Wang, Lili Chen, Xiaolin Zhang, Jiamao Li","doi":"10.3389/fnbot.2024.1395652","DOIUrl":"https://doi.org/10.3389/fnbot.2024.1395652","url":null,"abstract":"In Human-Robot Interaction (HRI), accurate 3D hand pose and mesh estimation hold critical importance. However, inferring reasonable and accurate poses in severe self-occlusion and high self-similarity remains an inherent challenge. In order to alleviate the ambiguity caused by invisible and similar joints during HRI, we propose a new Topology-aware Transformer network named HandGCNFormer with depth image as input, incorporating prior knowledge of hand kinematic topology into the network while modeling long-range contextual information. Specifically, we propose a novel Graphformer decoder with an additional Node-offset Graph Convolutional layer (NoffGConv). The Graphformer decoder optimizes the synergy between the Transformer and GCN, capturing long-range dependencies and local topological connections between joints. On top of that, we replace the standard MLP prediction head with a novel Topology-aware head to better exploit local topological constraints for more reasonable and accurate poses. Our method achieves state-of-the-art 3D hand pose estimation performance on four challenging datasets, including Hands2017, NYU, ICVL, and MSRA. To further demonstrate the effectiveness and scalability of our proposed Graphformer Decoder and Topology aware head, we extend our framework to HandGCNFormer-Mesh for the 3D hand mesh estimation task. The extended framework efficiently integrates a shape regressor with the original Graphformer Decoder and Topology aware head, producing Mano parameters. The results on the HO-3D dataset, which contains various and challenging occlusions, show that our HandGCNFormer-Mesh achieves competitive results compared to previous state-of-the-art 3D hand mesh estimation methods.","PeriodicalId":12628,"journal":{"name":"Frontiers in Neurorobotics","volume":null,"pages":null},"PeriodicalIF":3.1,"publicationDate":"2024-05-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140841369","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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 : 2024-05-01DOI: 10.3389/fnbot.2024.1376215
Tu Wang, Fujie Wang, Zhongye Xie, Feiyan Qin
In uncertain environments with robot input saturation, both model-based reinforcement learning (MBRL) and traditional controllers struggle to perform control tasks optimally. In this study, an algorithmic framework of Curiosity Model Policy Optimization (CMPO) is proposed by combining curiosity and model-based approach, where tracking errors are reduced via training agents on control gains for traditional model-free controllers. To begin with, a metric for judging positive and negative curiosity is proposed. Constrained optimization is employed to update the curiosity ratio, which improves the efficiency of agent training. Next, the novelty distance buffer ratio is defined to reduce bias between the environment and the model. Finally, CMPO is simulated with traditional controllers and baseline MBRL algorithms in the robotic environment designed with non-linear rewards. The experimental results illustrate that the algorithm achieves superior tracking performance and generalization capabilities.
{"title":"Curiosity model policy optimization for robotic manipulator tracking control with input saturation in uncertain environment","authors":"Tu Wang, Fujie Wang, Zhongye Xie, Feiyan Qin","doi":"10.3389/fnbot.2024.1376215","DOIUrl":"https://doi.org/10.3389/fnbot.2024.1376215","url":null,"abstract":"In uncertain environments with robot input saturation, both model-based reinforcement learning (MBRL) and traditional controllers struggle to perform control tasks optimally. In this study, an algorithmic framework of Curiosity Model Policy Optimization (CMPO) is proposed by combining curiosity and model-based approach, where tracking errors are reduced via training agents on control gains for traditional model-free controllers. To begin with, a metric for judging positive and negative curiosity is proposed. Constrained optimization is employed to update the curiosity ratio, which improves the efficiency of agent training. Next, the novelty distance buffer ratio is defined to reduce bias between the environment and the model. Finally, CMPO is simulated with traditional controllers and baseline MBRL algorithms in the robotic environment designed with non-linear rewards. The experimental results illustrate that the algorithm achieves superior tracking performance and generalization capabilities.","PeriodicalId":12628,"journal":{"name":"Frontiers in Neurorobotics","volume":null,"pages":null},"PeriodicalIF":3.1,"publicationDate":"2024-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140841364","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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 : 2024-05-01DOI: 10.3389/fnbot.2024.1342126
Huaqi Zhao, Xiang Peng, Su Wang, Jun-Bao Li, Jeng-Shyang Pan, Xiaoguang Su, Xiaomin Liu
The object detection method serves as the core technology within the unmanned driving perception module, extensively employed for detecting vehicles, pedestrians, traffic signs, and various objects. However, existing object detection methods still encounter three challenges in intricate unmanned driving scenarios: unsatisfactory performance in multi-scale object detection, inadequate accuracy in detecting small objects, and occurrences of false positives and missed detections in densely occluded environments. Therefore, this study proposes an improved object detection method for unmanned driving, leveraging Transformer architecture to address these challenges. First, a multi-scale Transformer feature extraction method integrated with channel attention is used to enhance the network's capability in extracting features across different scales. Second, a training method incorporating Query Denoising with Gaussian decay was employed to enhance the network's proficiency in learning representations of small objects. Third, a hybrid matching method combining Optimal Transport and Hungarian algorithms was used to facilitate the matching process between predicted and actual values, thereby enriching the network with more informative positive sample features. Experimental evaluations conducted on datasets including KITTI demonstrate that the proposed method achieves 3% higher mean Average Precision (mAP) than that of the existing methodologies.
{"title":"Improved object detection method for unmanned driving based on Transformers","authors":"Huaqi Zhao, Xiang Peng, Su Wang, Jun-Bao Li, Jeng-Shyang Pan, Xiaoguang Su, Xiaomin Liu","doi":"10.3389/fnbot.2024.1342126","DOIUrl":"https://doi.org/10.3389/fnbot.2024.1342126","url":null,"abstract":"The object detection method serves as the core technology within the unmanned driving perception module, extensively employed for detecting vehicles, pedestrians, traffic signs, and various objects. However, existing object detection methods still encounter three challenges in intricate unmanned driving scenarios: unsatisfactory performance in multi-scale object detection, inadequate accuracy in detecting small objects, and occurrences of false positives and missed detections in densely occluded environments. Therefore, this study proposes an improved object detection method for unmanned driving, leveraging Transformer architecture to address these challenges. First, a multi-scale Transformer feature extraction method integrated with channel attention is used to enhance the network's capability in extracting features across different scales. Second, a training method incorporating Query Denoising with Gaussian decay was employed to enhance the network's proficiency in learning representations of small objects. Third, a hybrid matching method combining Optimal Transport and Hungarian algorithms was used to facilitate the matching process between predicted and actual values, thereby enriching the network with more informative positive sample features. Experimental evaluations conducted on datasets including KITTI demonstrate that the proposed method achieves 3% higher mean Average Precision (mAP) than that of the existing methodologies.","PeriodicalId":12628,"journal":{"name":"Frontiers in Neurorobotics","volume":null,"pages":null},"PeriodicalIF":3.1,"publicationDate":"2024-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140841101","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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 : 2024-04-26DOI: 10.3389/fnbot.2024.1391247
Zhengchao Zhang, Lianke Zhou, Yuyang Wu, Nianbin Wang
IntroductionThe meta-learning methods have been widely used to solve the problem of few-shot learning. Generally, meta-learners are trained on a variety of tasks and then generalized to novel tasks.MethodsHowever, existing meta-learning methods do not consider the relationship between meta-tasks and novel tasks during the meta-training period, so that initial models of the meta-learner provide less useful meta-knowledge for the novel tasks. This leads to a weak generalization ability on novel tasks. Meanwhile, different initial models contain different meta-knowledge, which leads to certain differences in the learning effect of novel tasks during the meta-testing period. Therefore, this article puts forward a meta-optimization method based on situational meta-task construction and cooperation of multiple initial models. First, during the meta-training period, a method of constructing situational meta-task is proposed, and the selected candidate task sets provide more effective meta-knowledge for novel tasks. Then, during the meta-testing period, an ensemble model method based on meta-optimization is proposed to minimize the loss of inter-model cooperation in prediction, so that multiple models cooperation can realize the learning of novel tasks.ResultsThe above-mentioned methods are applied to popular few-shot character datasets and image recognition datasets. Furthermore, the experiment results indicate that the proposed method achieves good effects in few-shot classification tasks.DiscussionIn future work, we will extend our methods to provide more generalized and useful meta-knowledge to the model during the meta-training period when the novel few-shot tasks are completely invisible.
{"title":"The meta-learning method for the ensemble model based on situational meta-task","authors":"Zhengchao Zhang, Lianke Zhou, Yuyang Wu, Nianbin Wang","doi":"10.3389/fnbot.2024.1391247","DOIUrl":"https://doi.org/10.3389/fnbot.2024.1391247","url":null,"abstract":"IntroductionThe meta-learning methods have been widely used to solve the problem of few-shot learning. Generally, meta-learners are trained on a variety of tasks and then generalized to novel tasks.MethodsHowever, existing meta-learning methods do not consider the relationship between meta-tasks and novel tasks during the meta-training period, so that initial models of the meta-learner provide less useful meta-knowledge for the novel tasks. This leads to a weak generalization ability on novel tasks. Meanwhile, different initial models contain different meta-knowledge, which leads to certain differences in the learning effect of novel tasks during the meta-testing period. Therefore, this article puts forward a meta-optimization method based on situational meta-task construction and cooperation of multiple initial models. First, during the meta-training period, a method of constructing situational meta-task is proposed, and the selected candidate task sets provide more effective meta-knowledge for novel tasks. Then, during the meta-testing period, an ensemble model method based on meta-optimization is proposed to minimize the loss of inter-model cooperation in prediction, so that multiple models cooperation can realize the learning of novel tasks.ResultsThe above-mentioned methods are applied to popular few-shot character datasets and image recognition datasets. Furthermore, the experiment results indicate that the proposed method achieves good effects in few-shot classification tasks.DiscussionIn future work, we will extend our methods to provide more generalized and useful meta-knowledge to the model during the meta-training period when the novel few-shot tasks are completely invisible.","PeriodicalId":12628,"journal":{"name":"Frontiers in Neurorobotics","volume":null,"pages":null},"PeriodicalIF":3.1,"publicationDate":"2024-04-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140806225","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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 : 2024-04-23DOI: 10.3389/fnbot.2024.1401075
Yueguang Ge, Shaolin Zhang, Yinghao Cai, Tao Lu, Haitao Wang, Xiaolong Hui, Shuo Wang
IntroductionIn recent years, the perceptual capabilities of robots have been significantly enhanced. However, the task execution of the robots still lacks adaptive capabilities in unstructured and dynamic environments.MethodsIn this paper, we propose an ontology based autonomous robot task processing framework (ARTProF), to improve the robot's adaptability within unstructured and dynamic environments. ARTProF unifies ontological knowledge representation, reasoning, and autonomous task planning and execution into a single framework. The interface between the knowledge base and neural network-based object detection is first introduced in ARTProF to improve the robot's perception capabilities. A knowledge-driven manipulation operator based on Robot Operating System (ROS) is then designed to facilitate the interaction between the knowledge base and the robot's primitive actions. Additionally, an operation similarity model is proposed to endow the robot with the ability to generalize to novel objects. Finally, a dynamic task planning algorithm, leveraging ontological knowledge, equips the robot with adaptability to execute tasks in unstructured and dynamic environments.ResultsExperimental results on real-world scenarios and simulations demonstrate the effectiveness and efficiency of the proposed ARTProF framework.DiscussionIn future work, we will focus on refining the ARTProF framework by integrating neurosymbolic inference.
{"title":"Ontology based autonomous robot task processing framework","authors":"Yueguang Ge, Shaolin Zhang, Yinghao Cai, Tao Lu, Haitao Wang, Xiaolong Hui, Shuo Wang","doi":"10.3389/fnbot.2024.1401075","DOIUrl":"https://doi.org/10.3389/fnbot.2024.1401075","url":null,"abstract":"IntroductionIn recent years, the perceptual capabilities of robots have been significantly enhanced. However, the task execution of the robots still lacks adaptive capabilities in unstructured and dynamic environments.MethodsIn this paper, we propose an ontology based autonomous robot task processing framework (ARTProF), to improve the robot's adaptability within unstructured and dynamic environments. ARTProF unifies ontological knowledge representation, reasoning, and autonomous task planning and execution into a single framework. The interface between the knowledge base and neural network-based object detection is first introduced in ARTProF to improve the robot's perception capabilities. A knowledge-driven manipulation operator based on Robot Operating System (ROS) is then designed to facilitate the interaction between the knowledge base and the robot's primitive actions. Additionally, an operation similarity model is proposed to endow the robot with the ability to generalize to novel objects. Finally, a dynamic task planning algorithm, leveraging ontological knowledge, equips the robot with adaptability to execute tasks in unstructured and dynamic environments.ResultsExperimental results on real-world scenarios and simulations demonstrate the effectiveness and efficiency of the proposed ARTProF framework.DiscussionIn future work, we will focus on refining the ARTProF framework by integrating neurosymbolic inference.","PeriodicalId":12628,"journal":{"name":"Frontiers in Neurorobotics","volume":null,"pages":null},"PeriodicalIF":3.1,"publicationDate":"2024-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140885414","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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 : 2024-04-22DOI: 10.3389/fnbot.2024.1396979
Xudong Zhang, Yunlong Ge, Yifeng Wang, Jun Wang, Wenhao Wang, Lijun Lu
With the fast development of large-scale Photovoltaic (PV) plants, the automatic PV fault identification and positioning have become an important task for the PV intelligent systems, aiming to guarantee the safety, reliability, and productivity of large-scale PV plants. In this paper, we propose a residual learning-based robotic (UAV) image analysis model for low-voltage distributed PV fault identification and positioning. In our target scenario, the unmanned aerial vehicles (UAVs) are deployed to acquire moving images of low-voltage distributed PV power plants. To get desired robustness and accuracy of PV image detection, we integrate residual learning with attention mechanism into the UAV image analysis model based on you only look once v4 (YOLOv4) network. Then, we design the sophisticated multi-scale spatial pyramid fusion and use it to optimize the YOLOv4 network for the nuanced task of fault localization within PV arrays, where the Complete-IOU loss is incorporated in the predictive modeling phase, significantly enhancing the accuracy and efficiency of fault detection. A series of experimental comparisons in terms of the accuracy of fault positioning are conducted, and the experimental results verify the feasibility and effectiveness of the proposed model in dealing with the safety and reliability maintenance of low-voltage distributed PV systems.
{"title":"Residual learning-based robotic image analysis model for low-voltage distributed photovoltaic fault identification and positioning","authors":"Xudong Zhang, Yunlong Ge, Yifeng Wang, Jun Wang, Wenhao Wang, Lijun Lu","doi":"10.3389/fnbot.2024.1396979","DOIUrl":"https://doi.org/10.3389/fnbot.2024.1396979","url":null,"abstract":"With the fast development of large-scale Photovoltaic (PV) plants, the automatic PV fault identification and positioning have become an important task for the PV intelligent systems, aiming to guarantee the safety, reliability, and productivity of large-scale PV plants. In this paper, we propose a residual learning-based robotic (UAV) image analysis model for low-voltage distributed PV fault identification and positioning. In our target scenario, the unmanned aerial vehicles (UAVs) are deployed to acquire moving images of low-voltage distributed PV power plants. To get desired robustness and accuracy of PV image detection, we integrate residual learning with attention mechanism into the UAV image analysis model based on you only look once v4 (YOLOv4) network. Then, we design the sophisticated multi-scale spatial pyramid fusion and use it to optimize the YOLOv4 network for the nuanced task of fault localization within PV arrays, where the Complete-IOU loss is incorporated in the predictive modeling phase, significantly enhancing the accuracy and efficiency of fault detection. A series of experimental comparisons in terms of the accuracy of fault positioning are conducted, and the experimental results verify the feasibility and effectiveness of the proposed model in dealing with the safety and reliability maintenance of low-voltage distributed PV systems.","PeriodicalId":12628,"journal":{"name":"Frontiers in Neurorobotics","volume":null,"pages":null},"PeriodicalIF":3.1,"publicationDate":"2024-04-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140635325","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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 : 2024-04-10DOI: 10.3389/fnbot.2024.1364587
Feng Liu, Dongqi Li, Jian Gao
Multiagent Reinforcement Learning (MARL) has been well adopted due to its exceptional ability to solve multiagent decision-making problems. To further enhance learning efficiency, knowledge transfer algorithms have been developed, among which experience-sharing-based and action-advising-based transfer strategies share the mainstream. However, it is notable that, although there exist many successful applications of both strategies, they are not flawless. For the long-developed action-advising-based methods (namely KT-AA, short for knowledge transfer based on action advising), their data efficiency and scalability are not satisfactory. As for the newly proposed experience-sharing-based knowledge transfer methods (KT-ES), although the shortcomings of KT-AA have been partially overcome, they are incompetent to correct specific bad decisions in the later learning stage. To leverage the superiority of both KT-AA and KT-ES, this study proposes KT-Hybrid, a hybrid knowledge transfer approach. In the early learning phase, KT-ES methods are employed, expecting better data efficiency from KT-ES to enhance the policy to a basic level as soon as possible. Later, we focus on correcting specific errors made by the basic policy, trying to use KT-AA methods to further improve the performance. Simulations demonstrate that the proposed KT-Hybrid outperforms well-received action-advising- and experience-sharing-based methods.
{"title":"Hybrid knowledge transfer for MARL based on action advising and experience sharing","authors":"Feng Liu, Dongqi Li, Jian Gao","doi":"10.3389/fnbot.2024.1364587","DOIUrl":"https://doi.org/10.3389/fnbot.2024.1364587","url":null,"abstract":"Multiagent Reinforcement Learning (MARL) has been well adopted due to its exceptional ability to solve multiagent decision-making problems. To further enhance learning efficiency, knowledge transfer algorithms have been developed, among which experience-sharing-based and action-advising-based transfer strategies share the mainstream. However, it is notable that, although there exist many successful applications of both strategies, they are not flawless. For the long-developed action-advising-based methods (namely KT-AA, short for knowledge transfer based on action advising), their data efficiency and scalability are not satisfactory. As for the newly proposed experience-sharing-based knowledge transfer methods (KT-ES), although the shortcomings of KT-AA have been partially overcome, they are incompetent to correct specific bad decisions in the later learning stage. To leverage the superiority of both KT-AA and KT-ES, this study proposes KT-Hybrid, a hybrid knowledge transfer approach. In the early learning phase, KT-ES methods are employed, expecting better data efficiency from KT-ES to enhance the policy to a basic level as soon as possible. Later, we focus on correcting specific errors made by the basic policy, trying to use KT-AA methods to further improve the performance. Simulations demonstrate that the proposed KT-Hybrid outperforms well-received action-advising- and experience-sharing-based methods.","PeriodicalId":12628,"journal":{"name":"Frontiers in Neurorobotics","volume":null,"pages":null},"PeriodicalIF":3.1,"publicationDate":"2024-04-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140885486","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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 : 2024-04-09DOI: 10.3389/fnbot.2024.1397369
He Zhendong, Gao Xiangyang, Liu Zhiyuan, An Xiaoyu, Zheng Anping
Rail surface defects present a significant safety concern in railway operations. However, the scarcity of data poses challenges for employing deep learning in defect detection. This study proposes an enhanced ACGAN augmentation method to address these issues. Residual blocks mitigate vanishing gradient problems, while a spectral norm regularization-constrained discriminator improves stability and image quality. Substituting the generator’s deconvolution layer with upsampling and convolution operations enhances computational efficiency. A gradient penalty mechanism based on regret values addresses gradient abnormality concerns. Experimental validation demonstrates superior image clarity and classification accuracy compared to ACGAN, with a 17.6% reduction in FID value. MNIST dataset experiments verify the model’s generalization ability. This approach offers practical value for real-world applications.
{"title":"Rail surface defect data enhancement method based on improved ACGAN","authors":"He Zhendong, Gao Xiangyang, Liu Zhiyuan, An Xiaoyu, Zheng Anping","doi":"10.3389/fnbot.2024.1397369","DOIUrl":"https://doi.org/10.3389/fnbot.2024.1397369","url":null,"abstract":"Rail surface defects present a significant safety concern in railway operations. However, the scarcity of data poses challenges for employing deep learning in defect detection. This study proposes an enhanced ACGAN augmentation method to address these issues. Residual blocks mitigate vanishing gradient problems, while a spectral norm regularization-constrained discriminator improves stability and image quality. Substituting the generator’s deconvolution layer with upsampling and convolution operations enhances computational efficiency. A gradient penalty mechanism based on regret values addresses gradient abnormality concerns. Experimental validation demonstrates superior image clarity and classification accuracy compared to ACGAN, with a 17.6% reduction in FID value. MNIST dataset experiments verify the model’s generalization ability. This approach offers practical value for real-world applications.","PeriodicalId":12628,"journal":{"name":"Frontiers in Neurorobotics","volume":null,"pages":null},"PeriodicalIF":3.1,"publicationDate":"2024-04-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140584722","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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 : 2024-04-08DOI: 10.3389/fnbot.2024.1355170
Chuang Wang, Chupeng Su, Baozheng Sun, Gang Chen, Longhan Xie
IntroductionRobotic assembly tasks require precise manipulation and coordination, often necessitating advanced learning techniques to achieve efficient and effective performance. While residual reinforcement learning with a base policy has shown promise in this domain, existing base policy approaches often rely on hand-designed full-state features and policies or extensive demonstrations, limiting their applicability in semi-structured environments.MethodsIn this study, we propose an innovative Object-Embodiment-Centric Imitation and Residual Reinforcement Learning (OEC-IRRL) approach that leverages an object-embodiment-centric (OEC) task representation to integrate vision models with imitation and residual learning. By utilizing a single demonstration and minimizing interactions with the environment, our method aims to enhance learning efficiency and effectiveness. The proposed method involves three key steps: creating an object-embodiment-centric task representation, employing imitation learning for a base policy using via-point movement primitives for generalization to different settings, and utilizing residual RL for uncertainty-aware policy refinement during the assembly phase.ResultsThrough a series of comprehensive experiments, we investigate the impact of the OEC task representation on base and residual policy learning and demonstrate the effectiveness of the method in semi-structured environments. Our results indicate that the approach, requiring only a single demonstration and less than 1.2 h of interaction, improves success rates by 46% and reduces assembly time by 25%.DiscussionThis research presents a promising avenue for robotic assembly tasks, providing a viable solution without the need for specialized expertise or custom fixtures.
{"title":"Extended residual learning with one-shot imitation learning for robotic assembly in semi-structured environment","authors":"Chuang Wang, Chupeng Su, Baozheng Sun, Gang Chen, Longhan Xie","doi":"10.3389/fnbot.2024.1355170","DOIUrl":"https://doi.org/10.3389/fnbot.2024.1355170","url":null,"abstract":"IntroductionRobotic assembly tasks require precise manipulation and coordination, often necessitating advanced learning techniques to achieve efficient and effective performance. While residual reinforcement learning with a base policy has shown promise in this domain, existing base policy approaches often rely on hand-designed full-state features and policies or extensive demonstrations, limiting their applicability in semi-structured environments.MethodsIn this study, we propose an innovative Object-Embodiment-Centric Imitation and Residual Reinforcement Learning (OEC-IRRL) approach that leverages an object-embodiment-centric (OEC) task representation to integrate vision models with imitation and residual learning. By utilizing a single demonstration and minimizing interactions with the environment, our method aims to enhance learning efficiency and effectiveness. The proposed method involves three key steps: creating an object-embodiment-centric task representation, employing imitation learning for a base policy using via-point movement primitives for generalization to different settings, and utilizing residual RL for uncertainty-aware policy refinement during the assembly phase.ResultsThrough a series of comprehensive experiments, we investigate the impact of the OEC task representation on base and residual policy learning and demonstrate the effectiveness of the method in semi-structured environments. Our results indicate that the approach, requiring only a single demonstration and less than 1.2 h of interaction, improves success rates by 46% and reduces assembly time by 25%.DiscussionThis research presents a promising avenue for robotic assembly tasks, providing a viable solution without the need for specialized expertise or custom fixtures.","PeriodicalId":12628,"journal":{"name":"Frontiers in Neurorobotics","volume":null,"pages":null},"PeriodicalIF":3.1,"publicationDate":"2024-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140810365","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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：