Pub Date : 2025-01-23DOI: 10.1016/j.rcim.2025.102969
Wanyong Wang, Haohan Sun, Cong Chen, Ke Zhang
For composite plates side-sealing, traditional teaching-playback method is low-quality and inefficient, and cannot adapt to the rapid development of intelligent manufacturing. Aiming at this problem, an autonomous localization and welding path generation method based on binocular vision and lightweight deep learning network is proposed. Firstly, a lightweight background removal model based on VGG16-UNet (Visual Geometry Group Network-16 U-shaped Network) was proposed to eliminate different interference of illumination and redundant information. Secondly, Hough transform with RANSAC (Random Sample Consensus) correction was employed for accurate line extraction from unsharp workpiece edges. Then, an error compensation strategy was presented. Finally, a positioning accuracy of 0.47 mm was achieved, meeting the requirements for side-sealing. Autonomous localization and welding base path generation for composite plate billets with 20 mm depth grooves at a 3000 mm viewing distance were successfully realized. Welding results demonstrate that the proposed method is accurate and reliable, laying a solid foundation for further autonomous pass planning and adaptive controlling.
{"title":"Autonomous path generation for side-seal welding of composite plate billets based on binocular vision and lightweight network VGG16-UNet","authors":"Wanyong Wang, Haohan Sun, Cong Chen, Ke Zhang","doi":"10.1016/j.rcim.2025.102969","DOIUrl":"10.1016/j.rcim.2025.102969","url":null,"abstract":"<div><div>For composite plates side-sealing, traditional teaching-playback method is low-quality and inefficient, and cannot adapt to the rapid development of intelligent manufacturing. Aiming at this problem, an autonomous localization and welding path generation method based on binocular vision and lightweight deep learning network is proposed. Firstly, a lightweight background removal model based on VGG16-UNet (Visual Geometry Group Network-16 U-shaped Network) was proposed to eliminate different interference of illumination and redundant information. Secondly, Hough transform with RANSAC (Random Sample Consensus) correction was employed for accurate line extraction from unsharp workpiece edges. Then, an error compensation strategy was presented. Finally, a positioning accuracy of 0.47 mm was achieved, meeting the requirements for side-sealing. Autonomous localization and welding base path generation for composite plate billets with 20 mm depth grooves at a 3000 mm viewing distance were successfully realized. Welding results demonstrate that the proposed method is accurate and reliable, laying a solid foundation for further autonomous pass planning and adaptive controlling.</div></div>","PeriodicalId":21452,"journal":{"name":"Robotics and Computer-integrated Manufacturing","volume":"94 ","pages":"Article 102969"},"PeriodicalIF":9.1,"publicationDate":"2025-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143027367","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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 : 2025-01-23DOI: 10.1016/j.rcim.2025.102972
Teng Zhang , Fangyu Peng , Jianzhuang Wang , Zhao Yang , Xiaowei Tang , Rong Yan , Shengqiang Zhao , Runpeng Deng
In recent years, robotic machining has been widely noticed, especially in the manufacturing of large and complex parts, where large workspaces and flexible movements give it an even greater advantage. However, significant intrinsic errors, compliance errors due to weak stiffness of the joints, and spatially dependent nonlinear properties lead to significant challenges in high-precision machining. In this case, the dynamically changing contact area during the material removal process triggers a time-varying cutting force, which in combination with the characteristics of the robot body leads to a typical spatial–temporal coupling process that maps the error onto the workpiece. To address this process, an intelligent foreknowledge method for robot machining error with spatial–temporal feature coupling is proposed by considering the robot ontology error and the machining process. The proposed method carries out joint extraction of robot-related structured features and time-related serialized features and feature-level fusion mapping, respectively, and thus achieves accurate prediction of part machining errors. The proposed method is experimentally validated on eight inner wall workpieces of a cabin segment. Overall, the model achieved an optimal 0.026 mm RMSE on three test sub-workpieces. The ability of the proposed method to accurately extract spatial–temporal features and accurately predict machining errors is also verified through ablation experiments, parameter influence analysis experiments, and intermediate feature analysis. The proposed method takes data-driven as the core idea and spatial–temporal feature extraction as the dual perspective to achieve accurate prediction of robot machining error. It is of great significance for prediction-based accuracy compensation.
{"title":"Spatial–temporal feature fusion for intelligent foreknowledge of robotic machining errors","authors":"Teng Zhang , Fangyu Peng , Jianzhuang Wang , Zhao Yang , Xiaowei Tang , Rong Yan , Shengqiang Zhao , Runpeng Deng","doi":"10.1016/j.rcim.2025.102972","DOIUrl":"10.1016/j.rcim.2025.102972","url":null,"abstract":"<div><div>In recent years, robotic machining has been widely noticed, especially in the manufacturing of large and complex parts, where large workspaces and flexible movements give it an even greater advantage. However, significant intrinsic errors, compliance errors due to weak stiffness of the joints, and spatially dependent nonlinear properties lead to significant challenges in high-precision machining. In this case, the dynamically changing contact area during the material removal process triggers a time-varying cutting force, which in combination with the characteristics of the robot body leads to a typical spatial–temporal coupling process that maps the error onto the workpiece. To address this process, an intelligent foreknowledge method for robot machining error with spatial–temporal feature coupling is proposed by considering the robot ontology error and the machining process. The proposed method carries out joint extraction of robot-related structured features and time-related serialized features and feature-level fusion mapping, respectively, and thus achieves accurate prediction of part machining errors. The proposed method is experimentally validated on eight inner wall workpieces of a cabin segment. Overall, the model achieved an optimal 0.026 mm RMSE on three test sub-workpieces. The ability of the proposed method to accurately extract spatial–temporal features and accurately predict machining errors is also verified through ablation experiments, parameter influence analysis experiments, and intermediate feature analysis. The proposed method takes data-driven as the core idea and spatial–temporal feature extraction as the dual perspective to achieve accurate prediction of robot machining error. It is of great significance for prediction-based accuracy compensation.</div></div>","PeriodicalId":21452,"journal":{"name":"Robotics and Computer-integrated Manufacturing","volume":"94 ","pages":"Article 102972"},"PeriodicalIF":9.1,"publicationDate":"2025-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143027363","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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: 10.1016/j.rcim.2025.102957
Corentin Hubert , Nathan Odic , Marie Noel , Sidney Gharib , Seyedhossein H.H. Zargarbashi , Lama Séoud
The proliferation of tedious and repetitive tasks on production lines has accelerated the deployment of automated robots. This has also led to a demand for more flexible robots, known as cobots, that can work in collaboration with operators to perform a variety of tasks in different contexts. This paper explores the potential of computer vision-based hand gesture recognition as a means of human–robot interaction within cobotic platforms. Our research focuses on the challenges of gesture recognition in the face of visual occlusions and different camera viewpoints, typical of part finishing tasks in a real-world industrial setting. We introduce a new dataset, MuViH (Multi-View Hand gesture), which features a high variability in camera viewpoints, human operator characteristics, and occlusions, and is fully annotated for hand detection and gesture recognition. We then present a comprehensive hand gesture recognition pipeline that leverages this dataset. Our pipeline incorporates a multi-view aggregation step that significantly enhances gesture recognition accuracy, particularly in the case of visual occlusions. Thanks to extensive experiments and cross-validation on the MuViH dataset and another public dataset, HANDS, our approach demonstrates state-of-the-art performance in gesture recognition. This breakthrough underlines the potential of integrating robust vision-based interaction techniques into cobotic systems, improving flexibility and speed on the production line.
{"title":"MuViH: Multi-View Hand gesture dataset and recognition pipeline for human–robot interaction in a collaborative robotic finishing platform","authors":"Corentin Hubert , Nathan Odic , Marie Noel , Sidney Gharib , Seyedhossein H.H. Zargarbashi , Lama Séoud","doi":"10.1016/j.rcim.2025.102957","DOIUrl":"10.1016/j.rcim.2025.102957","url":null,"abstract":"<div><div>The proliferation of tedious and repetitive tasks on production lines has accelerated the deployment of automated robots. This has also led to a demand for more flexible robots, known as cobots, that can work in collaboration with operators to perform a variety of tasks in different contexts. This paper explores the potential of computer vision-based hand gesture recognition as a means of human–robot interaction within cobotic platforms. Our research focuses on the challenges of gesture recognition in the face of visual occlusions and different camera viewpoints, typical of part finishing tasks in a real-world industrial setting. We introduce a new dataset, MuViH (Multi-View Hand gesture), which features a high variability in camera viewpoints, human operator characteristics, and occlusions, and is fully annotated for hand detection and gesture recognition. We then present a comprehensive hand gesture recognition pipeline that leverages this dataset. Our pipeline incorporates a multi-view aggregation step that significantly enhances gesture recognition accuracy, particularly in the case of visual occlusions. Thanks to extensive experiments and cross-validation on the MuViH dataset and another public dataset, HANDS, our approach demonstrates state-of-the-art performance in gesture recognition. This breakthrough underlines the potential of integrating robust vision-based interaction techniques into cobotic systems, improving flexibility and speed on the production line.</div></div>","PeriodicalId":21452,"journal":{"name":"Robotics and Computer-integrated Manufacturing","volume":"94 ","pages":"Article 102957"},"PeriodicalIF":9.1,"publicationDate":"2025-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143027345","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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 : 2025-01-22DOI: 10.1016/j.rcim.2025.102968
Z.M. Bi , A. Mikkola , H. Handroos , C. Luo
With modularized architecture, integrated solutions can be configured by selecting and assembling a set of selected off-the-shelf functional modules to satisfy users’ needs optimally. While the attributes and properties of these modules are validated at components levels, the performances of system can be affected greatly by integration and interactions. Existing methodologies on system integration focus on system architecture, hardware and software reuses, communications, interfaces, and interoperation. There is the need to develop effective verification and validation (V&V) methods to assure the first-time-right from a virtual model to physical model in terms of the composability of system components to predict the performance of an integrated systems; note that not all attributes of composability can be verified by self-adaptability of cyber-physical systems. In this paper, we will focus on V&V of integrated robotic systems, and we will explore the relations of an integrated system with its components in terms of some performance criteria including functionalities, responsiveness, accuracy, and repeatability. The problem itself is newly formulated, and it is crucial for designers to predict and optimize system performance based on the selection and assemblage of system modules. The work in this paper opens new field of research in standardizing verification and validation process in designing collaborative robot systems
{"title":"A theoretical model to predict performance of integrated robotic systems","authors":"Z.M. Bi , A. Mikkola , H. Handroos , C. Luo","doi":"10.1016/j.rcim.2025.102968","DOIUrl":"10.1016/j.rcim.2025.102968","url":null,"abstract":"<div><div>With modularized architecture, integrated solutions can be configured by selecting and assembling a set of selected off-the-shelf functional modules to satisfy users’ needs optimally. While the attributes and properties of these modules are validated at components levels, the performances of system can be affected greatly by integration and interactions. Existing methodologies on system integration focus on system architecture, hardware and software reuses, communications, interfaces, and interoperation. There is the need to develop effective verification and validation (V&V) methods to assure the first-time-right from a virtual model to physical model in terms of the composability of system components to predict the performance of an integrated systems; note that not all attributes of composability can be verified by self-adaptability of cyber-physical systems. In this paper, we will focus on V&V of integrated robotic systems, and we will explore the relations of an integrated system with its components in terms of some performance criteria including functionalities, responsiveness, accuracy, and repeatability. The problem itself is newly formulated, and it is crucial for designers to predict and optimize system performance based on the selection and assemblage of system modules. The work in this paper opens new field of research in standardizing verification and validation process in designing collaborative robot systems</div></div>","PeriodicalId":21452,"journal":{"name":"Robotics and Computer-integrated Manufacturing","volume":"94 ","pages":"Article 102968"},"PeriodicalIF":9.1,"publicationDate":"2025-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143027343","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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: 10.1016/j.rcim.2025.102966
Yilin Mu, Lai Zou, Ziling Wang, Heng Li, Shengbo Yan, Wenxi Wang
Complex curvature changes and uneven allowance distribution significantly hinder the ability of traditional robotic belt grinding methods to achieve high-precision blade processing. To resolve this problem, a novel dynamic observer-based contact force control strategy is proposed in this paper by considering the dynamic contact force (DCF) model and partitioned force control (PFC) strategy. The DCF model is developed by considering the contact pressure distribution across different blade areas, while the over-grinding depth error is derived by analyzing the contact pressure coupling influenced by row spacing. The CC points with large allowance are divided into regions based on the variation of ideal normal contact force. Then, the reference normal contact force for each region is determined. Moreover, a dynamic observer-based adaptive impedance controller (DO-AIC) is developed to enhance reference normal contact force control. Verification experiment showed that DO-AIC increased force control accuracy by 78.27% compared to without the controller. Furthermore, four sets of robotic grinding experiments on turbine blades were performed to validate the superiority of the proposed method. The results showed that with DO-PFG, the surface profile accuracy at blade four areas improved to 0.244 mm, 0.188 mm, 0.193 mm, and 0.203 mm, representing improvements of 53.7%, 79.57%, 59.37%, and 67.26% compared to TG, respectively.
{"title":"A novel dynamic observer-based contact force control strategy in robotic grinding to improve blade profile accuracy","authors":"Yilin Mu, Lai Zou, Ziling Wang, Heng Li, Shengbo Yan, Wenxi Wang","doi":"10.1016/j.rcim.2025.102966","DOIUrl":"10.1016/j.rcim.2025.102966","url":null,"abstract":"<div><div>Complex curvature changes and uneven allowance distribution significantly hinder the ability of traditional robotic belt grinding methods to achieve high-precision blade processing. To resolve this problem, a novel dynamic observer-based contact force control strategy is proposed in this paper by considering the dynamic contact force (DCF) model and partitioned force control (PFC) strategy. The DCF model is developed by considering the contact pressure distribution across different blade areas, while the over-grinding depth error is derived by analyzing the contact pressure coupling influenced by row spacing. The CC points with large allowance are divided into regions based on the variation of ideal normal contact force. Then, the reference normal contact force for each region is determined. Moreover, a dynamic observer-based adaptive impedance controller (DO-AIC) is developed to enhance reference normal contact force control. Verification experiment showed that DO-AIC increased force control accuracy by 78.27% compared to without the controller. Furthermore, four sets of robotic grinding experiments on turbine blades were performed to validate the superiority of the proposed method. The results showed that with DO-PFG, the surface profile accuracy at blade four areas improved to 0.244 mm, 0.188 mm, 0.193 mm, and 0.203 mm, representing improvements of 53.7%, 79.57%, 59.37%, and 67.26% compared to TG, respectively.</div></div>","PeriodicalId":21452,"journal":{"name":"Robotics and Computer-integrated Manufacturing","volume":"94 ","pages":"Article 102966"},"PeriodicalIF":9.1,"publicationDate":"2025-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143027344","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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 : 2025-01-20DOI: 10.1016/j.rcim.2025.102964
Zhitao Gao , Chen Chen , Fangyu Peng , Yukui Zhang , Haoyan Liu , Wenke Zhou , Rong Yan , Xiaowei Tang
Collaborative robots are widely used in interaction tasks due to their low cost and high operational flexibility. However, compared to industrial robots, they have lower joint stiffness and are more sensitive to external environments, leading to larger motion tracking errors. Therefore, in interaction tasks within complex dynamic environments, such as wiping tasks with unexpected collision disturbances and drilling tasks with material property changes, maintaining the stability of the robot's motion velocity is crucial for improving task performance. To address these concerns, a comprehensive passive safety control framework is proposed in this work. The framework ensures system stability while imposing consistently constraints on non-passive power of the controller, resulting in high performance in the presence of external disturbances and material property changes. This is achieved by combining the Variable Energy Tank with the Adaptive Control Barrier Function method. On this basis, two key parameter design strategies of the framework are proposed, including a variable reference energy boundary strategy and an adaptive conservative factor strategy. The effectiveness of the proposed method is validated by real-world experiments involving wiping and drilling.
{"title":"Adaptive safety-critical control using a variable task energy tank for collaborative robot tasks under dynamic environments","authors":"Zhitao Gao , Chen Chen , Fangyu Peng , Yukui Zhang , Haoyan Liu , Wenke Zhou , Rong Yan , Xiaowei Tang","doi":"10.1016/j.rcim.2025.102964","DOIUrl":"10.1016/j.rcim.2025.102964","url":null,"abstract":"<div><div>Collaborative robots are widely used in interaction tasks due to their low cost and high operational flexibility. However, compared to industrial robots, they have lower joint stiffness and are more sensitive to external environments, leading to larger motion tracking errors. Therefore, in interaction tasks within complex dynamic environments, such as wiping tasks with unexpected collision disturbances and drilling tasks with material property changes, maintaining the stability of the robot's motion velocity is crucial for improving task performance. To address these concerns, a comprehensive passive safety control framework is proposed in this work. The framework ensures system stability while imposing consistently constraints on non-passive power of the controller, resulting in high performance in the presence of external disturbances and material property changes. This is achieved by combining the Variable Energy Tank with the Adaptive Control Barrier Function method. On this basis, two key parameter design strategies of the framework are proposed, including a variable reference energy boundary strategy and an adaptive conservative factor strategy. The effectiveness of the proposed method is validated by real-world experiments involving wiping and drilling.</div></div>","PeriodicalId":21452,"journal":{"name":"Robotics and Computer-integrated Manufacturing","volume":"94 ","pages":"Article 102964"},"PeriodicalIF":9.1,"publicationDate":"2025-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143027346","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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 : 2025-01-18DOI: 10.1016/j.rcim.2024.102946
Lin Ma , Ray Y. Zhong , Mingze Yuan , Kai Ding , Matthias Thürer , Yanghua Pan , Ting Qu , Geroge Q. Huang
Industry 5.0 emphasizes a human-centric concept, aiming to construct highly intelligent, sustainable, and resilient manufacturing systems. While a large body of literature has explored its concepts, architectures, enabling technologies, and practical applications, literature specifically focused on production planning and control solutions in industry 5.0 shops are scarce. Recent literature indicates that the well-being and skills of human workers significantly impact shop performance due to their highly variable activities and behaviors. Workload control has been recognized as a simple yet effective solution to mitigate the effects of high variability - both human and machine - through a three-layer filter for high-variety make-to-order shops, offering potential for Industry 5.0. However, the existing workload control concept has two significant limitations. First, it primarily focuses on the workload of machines while ignoring the potential impacts of humans, and; Second, this concept relied on the fixed processing times and lack flexibility to cope with changes in human subjective behaviors. In response, this study first presents a human-centric order release method based on workload control, enhancing its adaptability by considering uncertain human processing times. Furthermore, we introduce five shop floor priority dispatching rules to further investigate the potential impacts of additional factors on our proposed method. Simulation results show that the human-centric method outperforms the traditional machine-centric method, particularly in pure job shops. Meanwhile, when combining the human-centric order release method with the shop floor dispatching rules, the load-oriented dispatching rules significantly improve the shop's performance in terms of throughput time, while the time-oriented dispatching rules increase order delivery performance. Counterintuitively, integrating human-centric concept into the shop floor dispatching stage is noteworthy, i.e. human-centric shop floor dispatching rule. It does not enhance shop performance compared to the original dispatching rules, but rather deteriorates the performance of order release on most measures. The findings of this study have important implications for both research and practice in Industry 5.0.
{"title":"A human-centric order release method based on workload control in high-variety make-to-order shops towards Industry 5.0","authors":"Lin Ma , Ray Y. Zhong , Mingze Yuan , Kai Ding , Matthias Thürer , Yanghua Pan , Ting Qu , Geroge Q. Huang","doi":"10.1016/j.rcim.2024.102946","DOIUrl":"10.1016/j.rcim.2024.102946","url":null,"abstract":"<div><div>Industry 5.0 emphasizes a human-centric concept, aiming to construct highly intelligent, sustainable, and resilient manufacturing systems. While a large body of literature has explored its concepts, architectures, enabling technologies, and practical applications, literature specifically focused on production planning and control solutions in industry 5.0 shops are scarce. Recent literature indicates that the well-being and skills of human workers significantly impact shop performance due to their highly variable activities and behaviors. Workload control has been recognized as a simple yet effective solution to mitigate the effects of high variability - both human and machine - through a three-layer filter for high-variety make-to-order shops, offering potential for Industry 5.0. However, the existing workload control concept has two significant limitations. First, it primarily focuses on the workload of machines while ignoring the potential impacts of humans, and; Second, this concept relied on the fixed processing times and lack flexibility to cope with changes in human subjective behaviors. In response, this study first presents a human-centric order release method based on workload control, enhancing its adaptability by considering uncertain human processing times. Furthermore, we introduce five shop floor priority dispatching rules to further investigate the potential impacts of additional factors on our proposed method. Simulation results show that the human-centric method outperforms the traditional machine-centric method, particularly in pure job shops. Meanwhile, when combining the human-centric order release method with the shop floor dispatching rules, the load-oriented dispatching rules significantly improve the shop's performance in terms of throughput time, while the time-oriented dispatching rules increase order delivery performance. Counterintuitively, integrating human-centric concept into the shop floor dispatching stage is noteworthy, i.e. human-centric shop floor dispatching rule. It does not enhance shop performance compared to the original dispatching rules, but rather deteriorates the performance of order release on most measures. The findings of this study have important implications for both research and practice in Industry 5.0.</div></div>","PeriodicalId":21452,"journal":{"name":"Robotics and Computer-integrated Manufacturing","volume":"94 ","pages":"Article 102946"},"PeriodicalIF":9.1,"publicationDate":"2025-01-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142989081","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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 : 2025-01-18DOI: 10.1016/j.rcim.2025.102961
Zhaoyang Liao , Shufei Li , Fengyuan Xie , Guilin Yang , Xubin Lin , Zhihao Xu , Xuefeng Zhou
Traditional robotic gluing techniques suffer from uneven adhesive distribution and low coverage rates, particularly on complex surfaces and under varying process parameters, which impede their application in smart manufacturing. To overcome these limitations, this work presents a physical-simulation synergy approach for predicting glue line dimensions and optimizing toolpath planning, aimed at improving gluing quality and production efficiency. A predictive model is developed in the simulation layer using the Whale Optimization Algorithm combined with Gaussian Process Regression to accurately capture the nonlinear relationships between key process parameters and glue line dimensions. Building on this, a surrogate model is introduced to simulate glue line distribution after compression. To ensure full coverage and high uniformity, a high-uniformity toolpath planning strategy is implemented, utilizing growth-based Hilbert curves and conformal mapping to generate efficient gluing toolpaths on complex surfaces in physical environments. Experimental results validate the effectiveness of the proposed method in accurately predicting glue dimensions, enhancing coverage, and improving adhesive performance, demonstrating its suitability for applications involving complex surface geometries.
{"title":"A Physical-simulation synergy approach for high-uniformity robotic gluing","authors":"Zhaoyang Liao , Shufei Li , Fengyuan Xie , Guilin Yang , Xubin Lin , Zhihao Xu , Xuefeng Zhou","doi":"10.1016/j.rcim.2025.102961","DOIUrl":"10.1016/j.rcim.2025.102961","url":null,"abstract":"<div><div>Traditional robotic gluing techniques suffer from uneven adhesive distribution and low coverage rates, particularly on complex surfaces and under varying process parameters, which impede their application in smart manufacturing. To overcome these limitations, this work presents a physical-simulation synergy approach for predicting glue line dimensions and optimizing toolpath planning, aimed at improving gluing quality and production efficiency. A predictive model is developed in the simulation layer using the Whale Optimization Algorithm combined with Gaussian Process Regression to accurately capture the nonlinear relationships between key process parameters and glue line dimensions. Building on this, a surrogate model is introduced to simulate glue line distribution after compression. To ensure full coverage and high uniformity, a high-uniformity toolpath planning strategy is implemented, utilizing growth-based Hilbert curves and conformal mapping to generate efficient gluing toolpaths on complex surfaces in physical environments. Experimental results validate the effectiveness of the proposed method in accurately predicting glue dimensions, enhancing coverage, and improving adhesive performance, demonstrating its suitability for applications involving complex surface geometries.</div></div>","PeriodicalId":21452,"journal":{"name":"Robotics and Computer-integrated Manufacturing","volume":"94 ","pages":"Article 102961"},"PeriodicalIF":9.1,"publicationDate":"2025-01-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142989080","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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: 10.1016/j.rcim.2025.102959
Xiaohan Wang , Lin Zhang , Lihui Wang , Enrique Ruiz Zuñiga , Xi Vincent Wang , Erik Flores-García
Dynamic order picking has usually demonstrated significant impacts on production efficiency in warehouse management. In the context of an automotive-part warehouse, this paper addresses a dynamic multi-tour order-picking problem based on a novel attention-aware deep reinforcement learning-based (ADRL) method. The multi-tour represents that one order-picking task must be split into multiple tours due to the cart capacity and the operator’s workload constraints. First, the multi-tour order-picking problem is formulated as a mathematical model, and then reformulated as a Markov decision process. Second, a novel DRL-based method is proposed to solve it effectively. Compared to the existing DRL-based methods, this approach employs multi-head attention to perceive warehouse situations. Additionally, three improvements are proposed to further strengthen the solution quality and generalization, including (1) the extra location representation to align the batch length during training, (2) the dynamic decoding to integrate real-time information of the warehouse environment during inference, and (3) the proximal policy optimization with entropy bonus to facilitate action exploration during training. Finally, comparison experiments based on thousands of order-picking instances from the Swedish warehouse validated that the proposed ADRL could outperform the other twelve DRL-based methods at most by 40.6%, considering the optimization objective. Furthermore, the performance gap between ADRL and seven evolutionary algorithms is controlled within 3%, while ADRL can be hundreds or thousands of times faster than these EAs regarding the solving speed.
{"title":"Dynamic multi-tour order picking in an automotive-part warehouse based on attention-aware deep reinforcement learning","authors":"Xiaohan Wang , Lin Zhang , Lihui Wang , Enrique Ruiz Zuñiga , Xi Vincent Wang , Erik Flores-García","doi":"10.1016/j.rcim.2025.102959","DOIUrl":"10.1016/j.rcim.2025.102959","url":null,"abstract":"<div><div>Dynamic order picking has usually demonstrated significant impacts on production efficiency in warehouse management. In the context of an automotive-part warehouse, this paper addresses a dynamic multi-tour order-picking problem based on a novel attention-aware deep reinforcement learning-based (ADRL) method. The multi-tour represents that one order-picking task must be split into multiple tours due to the cart capacity and the operator’s workload constraints. First, the multi-tour order-picking problem is formulated as a mathematical model, and then reformulated as a Markov decision process. Second, a novel DRL-based method is proposed to solve it effectively. Compared to the existing DRL-based methods, this approach employs multi-head attention to perceive warehouse situations. Additionally, three improvements are proposed to further strengthen the solution quality and generalization, including (1) the extra location representation to align the batch length during training, (2) the dynamic decoding to integrate real-time information of the warehouse environment during inference, and (3) the proximal policy optimization with entropy bonus to facilitate action exploration during training. Finally, comparison experiments based on thousands of order-picking instances from the Swedish warehouse validated that the proposed ADRL could outperform the other twelve DRL-based methods at most by 40.6%, considering the optimization objective. Furthermore, the performance gap between ADRL and seven evolutionary algorithms is controlled within 3%, while ADRL can be hundreds or thousands of times faster than these EAs regarding the solving speed.</div></div>","PeriodicalId":21452,"journal":{"name":"Robotics and Computer-integrated Manufacturing","volume":"94 ","pages":"Article 102959"},"PeriodicalIF":9.1,"publicationDate":"2025-01-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142989085","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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 : 2025-01-15DOI: 10.1016/j.rcim.2025.102962
Zhipeng Ma , Ming Zhao , Xuebin Dai , Yang Chen
Process monitoring in high-speed machining (HSM) is essential to guarantee product quality and improve manufacturing efficiency. Nevertheless, the data acquired from practical machining processes are completely unlabeled and severely unbalanced, which may be seriously insufficient to support deep learning-based anomaly detection. Furthermore, the collected signals are inevitably contaminated by environmental noises and uncertain factors. How to remove these disturbances according to data distribution characteristics remains a challenging issue. To tackle these limitations, a novel interpretable machine learning approach, called hybrid regularized support vector data description (H-SVDD), is proposed for unsupervised anomaly detection during HSM. In this work, an adaptive local kernel density estimate is first constructed to eliminate outlier interferences, and assigns interpretable weights to optimize the SVDD for improving detection accuracy. Subsequently, by introducing the lp-norm penalty mechanism, a generalized probability density regularized SVDD is innovatively designed to enhance the descriptive capability for complex machining processes. Finally, a hyperparameter tuning strategy based on Bayesian optimization is developed to improve generalizability and stability. The data collected from CNC machines are used to verify the superiority of the proposed method. Experimental results show that the proposed H-SVDD has higher detection accuracy than current SVDD methods and eliminates false alarms caused by noise interferences. This work may provide a useful solution for independently perceiving the health conditions of HSM.
{"title":"Anomaly detection for high-speed machining using hybrid regularized support vector data description","authors":"Zhipeng Ma , Ming Zhao , Xuebin Dai , Yang Chen","doi":"10.1016/j.rcim.2025.102962","DOIUrl":"10.1016/j.rcim.2025.102962","url":null,"abstract":"<div><div>Process monitoring in high-speed machining (HSM) is essential to guarantee product quality and improve manufacturing efficiency. Nevertheless, the data acquired from practical machining processes are completely unlabeled and severely unbalanced, which may be seriously insufficient to support deep learning-based anomaly detection. Furthermore, the collected signals are inevitably contaminated by environmental noises and uncertain factors. How to remove these disturbances according to data distribution characteristics remains a challenging issue. To tackle these limitations, a novel interpretable machine learning approach, called hybrid regularized support vector data description (H-SVDD), is proposed for unsupervised anomaly detection during HSM. In this work, an adaptive local kernel density estimate is first constructed to eliminate outlier interferences, and assigns interpretable weights to optimize the SVDD for improving detection accuracy. Subsequently, by introducing the <em>l<sub>p</sub></em>-norm penalty mechanism, a generalized probability density regularized SVDD is innovatively designed to enhance the descriptive capability for complex machining processes. Finally, a hyperparameter tuning strategy based on Bayesian optimization is developed to improve generalizability and stability. The data collected from CNC machines are used to verify the superiority of the proposed method. Experimental results show that the proposed H-SVDD has higher detection accuracy than current SVDD methods and eliminates false alarms caused by noise interferences. This work may provide a useful solution for independently perceiving the health conditions of HSM.</div></div>","PeriodicalId":21452,"journal":{"name":"Robotics and Computer-integrated Manufacturing","volume":"94 ","pages":"Article 102962"},"PeriodicalIF":9.1,"publicationDate":"2025-01-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142989526","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号