Chao Lin, Shuhei Abe, Shitao Zheng, Xianfeng Li, Pang-jo Chun
Focusing on learning-based semantic segmentation (SS) methods for bridge point cloud data (PCD), this study proposes a structure-oriented concept (SOC) with training focused on the spatial distribution patterns of bridge components, including both the horizontally absolute location of each component and its vertically relative position compared with other components. Then a structure-oriented loss (SOL) function, which embodies the core of SOC, is defined accordingly, and it is compared to five cutting-edge loss functions on a collected bridge PCD dataset. In contrast to the limitations of other loss functions, SOL significantly improves the overall evaluation metrics of overall accuracy (6.53%) and mean intersection over union (mean IoU: 8.67%). The IoU of the category “others” is improved by 8.44%, which is very important for automating the time-consuming denoising process. Furthermore, the demonstrated robustness of SOC and SOL reveal great potential to improve the performance of other SS models.
{"title":"A structure-oriented loss function for automated semantic segmentation of bridge point clouds","authors":"Chao Lin, Shuhei Abe, Shitao Zheng, Xianfeng Li, Pang-jo Chun","doi":"10.1111/mice.13422","DOIUrl":"https://doi.org/10.1111/mice.13422","url":null,"abstract":"Focusing on learning-based semantic segmentation (SS) methods for bridge point cloud data (PCD), this study proposes a structure-oriented concept (SOC) with training focused on the spatial distribution patterns of bridge components, including both the horizontally absolute location of each component and its vertically relative position compared with other components. Then a structure-oriented loss (SOL) function, which embodies the core of SOC, is defined accordingly, and it is compared to five cutting-edge loss functions on a collected bridge PCD dataset. In contrast to the limitations of other loss functions, SOL significantly improves the overall evaluation metrics of overall accuracy (6.53%) and mean intersection over union (mean IoU: 8.67%). The IoU of the category “others” is improved by 8.44%, which is very important for automating the time-consuming denoising process. Furthermore, the demonstrated robustness of SOC and SOL reveal great potential to improve the performance of other SS models.","PeriodicalId":156,"journal":{"name":"Computer-Aided Civil and Infrastructure Engineering","volume":"9 1","pages":""},"PeriodicalIF":11.775,"publicationDate":"2025-01-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142967875","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In sewer pipe closed‐circuit television inspection, accurate temporal defect localization is essential for effective pipe assessment. Industry standards typically do not require time interval annotations, which are more informative but lead to additional costs for fully supervised methods. Additionally, differences in scene types and camera motion patterns between pipe inspections and temporal action localization (TAL) hinder the effective transfer of point‐supervised TAL methods. Therefore, this study presents a semi‐supervised multi‐prototype‐based method incorporating visual odometry for enhanced attention guidance (PipeSPO). The semi‐supervised multi‐prototype‐based method effectively leverages both unlabeled data and time‐point annotations, which enhances performance and reduces annotation costs. Meanwhile, visual odometry features exploit the camera's unique motion patterns in pipe videos, offering additional insights to inform the model. Experiments on real‐world datasets demonstrate that PipeSPO achieves 41.89% AP across intersection over union thresholds of 0.1–0.7, improving by 8.14% over current state‐of‐the‐art methods.
{"title":"Semi‐supervised pipe video temporal defect interval localization","authors":"Zhu Huang, Gang Pan, Chao Kang, YaoZhi Lv","doi":"10.1111/mice.13403","DOIUrl":"https://doi.org/10.1111/mice.13403","url":null,"abstract":"In sewer pipe closed‐circuit television inspection, accurate temporal defect localization is essential for effective pipe assessment. Industry standards typically do not require time interval annotations, which are more informative but lead to additional costs for fully supervised methods. Additionally, differences in scene types and camera motion patterns between pipe inspections and temporal action localization (TAL) hinder the effective transfer of point‐supervised TAL methods. Therefore, this study presents a semi‐supervised multi‐prototype‐based method incorporating visual odometry for enhanced attention guidance (PipeSPO). The semi‐supervised multi‐prototype‐based method effectively leverages both unlabeled data and time‐point annotations, which enhances performance and reduces annotation costs. Meanwhile, visual odometry features exploit the camera's unique motion patterns in pipe videos, offering additional insights to inform the model. Experiments on real‐world datasets demonstrate that PipeSPO achieves 41.89% AP across intersection over union thresholds of 0.1–0.7, improving by 8.14% over current state‐of‐the‐art methods.","PeriodicalId":156,"journal":{"name":"Computer-Aided Civil and Infrastructure Engineering","volume":"24 1","pages":""},"PeriodicalIF":11.775,"publicationDate":"2025-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142939987","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Xiaowei Liu, Jinqu Chen, Bo Du, Xu Yan, Qiyuan Peng, Jun Shen
Unlike most urban rail transit (URT) resilience studies on URT lines or networks under major disturbances, this paper focuses on the resilience assessment of URT stations under high-frequency daily disturbances with minor impacts. A resilience assessment metric with different resilience levels is proposed, which is calculated based on multiple criteria, including the number of delayed passengers, degree of congestion, economic loss from service suppliers’ perspective, extra in-station travel time, extra walking distance, and extra waiting time from passengers’ perspective. A two-stage passenger flow redistribution model is developed with stage one focusing on route adjustment under disturbance, while stage two determining the walking path within the disrupted station. A case study of Simaqiao Station in the Chengdu subway network in China is conducted. The numerical results indicate that this station demonstrates strong resilience in most scenarios, although it faces challenges under certain identified disturbances.
{"title":"Resilience assessment of urban rail transit stations considering disturbance and time-varying passenger flow","authors":"Xiaowei Liu, Jinqu Chen, Bo Du, Xu Yan, Qiyuan Peng, Jun Shen","doi":"10.1111/mice.13400","DOIUrl":"https://doi.org/10.1111/mice.13400","url":null,"abstract":"Unlike most urban rail transit (URT) resilience studies on URT lines or networks under major disturbances, this paper focuses on the resilience assessment of URT stations under high-frequency daily disturbances with minor impacts. A resilience assessment metric with different resilience levels is proposed, which is calculated based on multiple criteria, including the number of delayed passengers, degree of congestion, economic loss from service suppliers’ perspective, extra in-station travel time, extra walking distance, and extra waiting time from passengers’ perspective. A two-stage passenger flow redistribution model is developed with stage one focusing on route adjustment under disturbance, while stage two determining the walking path within the disrupted station. A case study of Simaqiao Station in the Chengdu subway network in China is conducted. The numerical results indicate that this station demonstrates strong resilience in most scenarios, although it faces challenges under certain identified disturbances.","PeriodicalId":156,"journal":{"name":"Computer-Aided Civil and Infrastructure Engineering","volume":"132 1","pages":""},"PeriodicalIF":11.775,"publicationDate":"2025-01-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142936551","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Ruru Hao, Shixiao Liang, Ziyang Zhai, Hang Zhou, Xin Wang, Xiaopeng Li, Tianhao Guan
The deployment of sensors to monitor traffic flow between origin–destination (OD) pairs, within a specified budget, remains a critical concern for both academic researchers and transportation managers. While these technologies are essential for capturing traffic data, the aspect of privacy has often been overlooked. To bridge this gap, this paper introduced the concept of privacy distance and then proposed an integer programming model to optimize traffic sensor locations by maximizing the coverage of traffic flow while taking into account the punishment brought by the risk of privacy leakage. Furthermore, to address the computational efficiency problem in large-scale networks, a flow threshold is set to properly remove some OD pairs to balance the model tractability and computational efficiency. Two case studies of different sizes are carried out to discuss the performance. Case 1 validated the effectiveness of the model, while case 2 demonstrated its capability to handle large-scale problems. The experimental results show that for large-scale networks, setting a flow threshold can reduce computation time by 96% at the cost of sacrificing 12% of the OD coverage.
{"title":"Privacy-preserving awareness in sensor deployment for traffic flow surveillance","authors":"Ruru Hao, Shixiao Liang, Ziyang Zhai, Hang Zhou, Xin Wang, Xiaopeng Li, Tianhao Guan","doi":"10.1111/mice.13418","DOIUrl":"https://doi.org/10.1111/mice.13418","url":null,"abstract":"The deployment of sensors to monitor traffic flow between origin–destination (OD) pairs, within a specified budget, remains a critical concern for both academic researchers and transportation managers. While these technologies are essential for capturing traffic data, the aspect of privacy has often been overlooked. To bridge this gap, this paper introduced the concept of privacy distance and then proposed an integer programming model to optimize traffic sensor locations by maximizing the coverage of traffic flow while taking into account the punishment brought by the risk of privacy leakage. Furthermore, to address the computational efficiency problem in large-scale networks, a flow threshold is set to properly remove some OD pairs to balance the model tractability and computational efficiency. Two case studies of different sizes are carried out to discuss the performance. Case 1 validated the effectiveness of the model, while case 2 demonstrated its capability to handle large-scale problems. The experimental results show that for large-scale networks, setting a flow threshold can reduce computation time by 96% at the cost of sacrificing 12% of the OD coverage.","PeriodicalId":156,"journal":{"name":"Computer-Aided Civil and Infrastructure Engineering","volume":"23 1","pages":""},"PeriodicalIF":11.775,"publicationDate":"2025-01-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142936550","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Alireza Ghiasi, Zhen Zhang, Zijie Zeng, Ching Tai Ng, Abdul Hamid Sheikh, Javen Qinfeng Shi
Corrosion is one of the main damages in steel bridges, which appears as a loss of material and sectional area and causes member failure over time. A reliable bridge management system not only should help in preventing catastrophic structural failure by employing an in-time anomaly detection approach for all the bridges within a network but also should reduce overall network costs commonly raised by expensive inspections. This paper proposes a deep learning approach to generalize anomaly detection due to section losses in steel bridges based on Siamese convolutional neural network (SCNN). A series of steel beams and bridges with various cross-sections and lengths are considered to examine the performance of SCNN in generalizing anomaly detection in these structures. The study considered data from finite element simulations and experiments. The results reveal that the proposed integrated SCNN can detect anomalies successfully according to Australian standard AS7636 with reasonably high accuracy.
{"title":"Generalization of anomaly detection in bridge structures using a vibration-based Siamese convolutional neural network","authors":"Alireza Ghiasi, Zhen Zhang, Zijie Zeng, Ching Tai Ng, Abdul Hamid Sheikh, Javen Qinfeng Shi","doi":"10.1111/mice.13411","DOIUrl":"https://doi.org/10.1111/mice.13411","url":null,"abstract":"Corrosion is one of the main damages in steel bridges, which appears as a loss of material and sectional area and causes member failure over time. A reliable bridge management system not only should help in preventing catastrophic structural failure by employing an in-time anomaly detection approach for all the bridges within a network but also should reduce overall network costs commonly raised by expensive inspections. This paper proposes a deep learning approach to generalize anomaly detection due to section losses in steel bridges based on Siamese convolutional neural network (SCNN). A series of steel beams and bridges with various cross-sections and lengths are considered to examine the performance of SCNN in generalizing anomaly detection in these structures. The study considered data from finite element simulations and experiments. The results reveal that the proposed integrated SCNN can detect anomalies successfully according to Australian standard AS7636 with reasonably high accuracy.","PeriodicalId":156,"journal":{"name":"Computer-Aided Civil and Infrastructure Engineering","volume":"67 1","pages":""},"PeriodicalIF":11.775,"publicationDate":"2025-01-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142936545","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The cover image is based on the article A rendering-based lightweight network for segmentation of high-resolution crack images by Weiwei Chen et al., https://doi.org/10.1111/mice.13290.
{"title":"Cover Image, Volume 40, Issue 3","authors":"","doi":"10.1111/mice.13416","DOIUrl":"https://doi.org/10.1111/mice.13416","url":null,"abstract":"<b>The cover image</b> is based on the article <i>A rendering-based lightweight network for segmentation of high-resolution crack images</i> by Weiwei Chen et al., https://doi.org/10.1111/mice.13290.","PeriodicalId":156,"journal":{"name":"Computer-Aided Civil and Infrastructure Engineering","volume":"82 1","pages":""},"PeriodicalIF":11.775,"publicationDate":"2025-01-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142935159","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Three-dimensional (3D) buried object detection using ground penetrating radar (GPR) benefits from the powerful capacity of image-wise deep neural networks. However, it still faces the challenge of information loss from raw GPR signals to two- and three-dimensional images, such as the frequency-domain information loss when normalizing GPR signals into gray-scale images and spatial information loss when using stacked B- and C-scan images to replace raw GPR signals as inputs. To solve the challenge, this study has proposed an ENNreg-transformer model, directly using raw 3D GPR signals to perform buried object detection. In the proposed model, 3D GPR signals are first converted into sequential voxelization to obtain spatiotemporal features. The features are then aggregated by an intuition-guided feature aggregation layer to simulate the expert behavior to analyze 3D GPR data. Finally, an evidential detection header outputs 3D interval-based bounding boxes for buried object detection. The experiment on two 3D GPR road datasets demonstrates that the proposed model exceeds other state-of-the-art models on the tasks thanks to raw 3D signals and intuition-guided feature aggregation. In addition, the interval-based bounding box represents the spatial bounding-box uncertainty, which derives from the inherent limitations of GPR and deep networks.
{"title":"Evidential transformer for buried object detection in ground penetrating radar signals and interval-based bounding box","authors":"Zheng Tong, Yiming Zhang, Tao Ma","doi":"10.1111/mice.13417","DOIUrl":"https://doi.org/10.1111/mice.13417","url":null,"abstract":"Three-dimensional (3D) buried object detection using ground penetrating radar (GPR) benefits from the powerful capacity of image-wise deep neural networks. However, it still faces the challenge of information loss from raw GPR signals to two- and three-dimensional images, such as the frequency-domain information loss when normalizing GPR signals into gray-scale images and spatial information loss when using stacked B- and C-scan images to replace raw GPR signals as inputs. To solve the challenge, this study has proposed an ENNreg-transformer model, directly using raw 3D GPR signals to perform buried object detection. In the proposed model, 3D GPR signals are first converted into sequential voxelization to obtain spatiotemporal features. The features are then aggregated by an intuition-guided feature aggregation layer to simulate the expert behavior to analyze 3D GPR data. Finally, an evidential detection header outputs 3D interval-based bounding boxes for buried object detection. The experiment on two 3D GPR road datasets demonstrates that the proposed model exceeds other state-of-the-art models on the tasks thanks to raw 3D signals and intuition-guided feature aggregation. In addition, the interval-based bounding box represents the spatial bounding-box uncertainty, which derives from the inherent limitations of GPR and deep networks.","PeriodicalId":156,"journal":{"name":"Computer-Aided Civil and Infrastructure Engineering","volume":"23 1","pages":""},"PeriodicalIF":11.775,"publicationDate":"2025-01-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142936544","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The longevity of porous asphalt pavement is inevitably compromised by the clogging of voids by various particles, leading to a degradation in its drainage function. Numerical simulations with real pore structures were used to investigate the clogging behavior of porous asphalt concrete (PAC) to clearly and intuitively understand its void clogging process. In this study, a three-dimensional model of the real void was created by computed tomography scanning. The change before and after void clogging of PAC was characterized by seepage pressure and seepage velocity in the seepage field. The computational fluid dynamics-discrete element method coupling method was used to visually describe the dynamic evolution of clogging particles in porous asphalt voids. Findings reveal that the most influential particle size for clogging in PAC-13 with 18% and 20% porosity ranged between 0.15 and 0.6 mm. In contrast, for PAC-13 with 25% porosity, the sensitive size was 0.3–1.18 mm. When clogging occurred, large particles predominantly obstructed the void inlets, prompting a refinement in the void structure. Subsequent particles either traversed the void, accumulating at the entrances of finer voids, or filled up progressively, leading to eventual clogging. Small particles either exited directly through the voids or accumulated in the bends of the voids, making the voids clogged directly. Consequently, the clogging behavior of porous asphalt was classified into three types: surface-filling clogging, void refining filter clogging, and void bending or semi-connecting clogging. These findings provide a scientific basis for optimizing PAC design and developing conservation strategies.
{"title":"Evolution of clogging of porous asphalt concrete in the seepage process through integration of computer tomography, computational fluid dynamics, and discrete element method","authors":"Bo Li, Yunpeng Zhang, Dingbang Wei, Tengfei Yao, Yongping Hu, Hui Dou","doi":"10.1111/mice.13419","DOIUrl":"https://doi.org/10.1111/mice.13419","url":null,"abstract":"The longevity of porous asphalt pavement is inevitably compromised by the clogging of voids by various particles, leading to a degradation in its drainage function. Numerical simulations with real pore structures were used to investigate the clogging behavior of porous asphalt concrete (PAC) to clearly and intuitively understand its void clogging process. In this study, a three-dimensional model of the real void was created by computed tomography scanning. The change before and after void clogging of PAC was characterized by seepage pressure and seepage velocity in the seepage field. The computational fluid dynamics-discrete element method coupling method was used to visually describe the dynamic evolution of clogging particles in porous asphalt voids. Findings reveal that the most influential particle size for clogging in PAC-13 with 18% and 20% porosity ranged between 0.15 and 0.6 mm. In contrast, for PAC-13 with 25% porosity, the sensitive size was 0.3–1.18 mm. When clogging occurred, large particles predominantly obstructed the void inlets, prompting a refinement in the void structure. Subsequent particles either traversed the void, accumulating at the entrances of finer voids, or filled up progressively, leading to eventual clogging. Small particles either exited directly through the voids or accumulated in the bends of the voids, making the voids clogged directly. Consequently, the clogging behavior of porous asphalt was classified into three types: surface-filling clogging, void refining filter clogging, and void bending or semi-connecting clogging. These findings provide a scientific basis for optimizing PAC design and developing conservation strategies.","PeriodicalId":156,"journal":{"name":"Computer-Aided Civil and Infrastructure Engineering","volume":"3 1","pages":""},"PeriodicalIF":11.775,"publicationDate":"2025-01-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142936549","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Jifu Guo, Shengguang Bai, Xun Li, Kai Xian, Erjian Liu, Wenting Ding, Xizhi Ma
This study primarily focuses on generating mobility flow in regions and cities, which plays an important role in urban planning and management. The majority of existing mobility flow models, including conventional statistical models and deep learning-based models, are heavily dependent on historical data to predict future mobility flows. The application of these models poses significant challenges in the planning and construction of emerging cities and regions, particularly in developing countries experiencing swift urbanization. These challenges are exacerbated by a dearth of historical data and rapid shifts in mobility patterns. Consequently, the scenario necessitates a mobility flow generation model capable of generating flows without historical data. This study introduces the universal geography neural network, an algorithm designed to glean potential patterns in human mobility across diverse cities and temporal spans. This is achieved through the analysis of substantial quantities of location-based data, resulting in the generation of mobility flows within a city. Our experiment, designed to extract various features and generate fine-grained mobility flows in the testing set, outperforms both traditional models and state-of-the-art deep learning models. Moreover, our model has proven capable of generating reliable results across various time periods and grid areas.
{"title":"A universal geography neural network for mobility flow prediction in planning scenarios","authors":"Jifu Guo, Shengguang Bai, Xun Li, Kai Xian, Erjian Liu, Wenting Ding, Xizhi Ma","doi":"10.1111/mice.13398","DOIUrl":"https://doi.org/10.1111/mice.13398","url":null,"abstract":"This study primarily focuses on generating mobility flow in regions and cities, which plays an important role in urban planning and management. The majority of existing mobility flow models, including conventional statistical models and deep learning-based models, are heavily dependent on historical data to predict future mobility flows. The application of these models poses significant challenges in the planning and construction of emerging cities and regions, particularly in developing countries experiencing swift urbanization. These challenges are exacerbated by a dearth of historical data and rapid shifts in mobility patterns. Consequently, the scenario necessitates a mobility flow generation model capable of generating flows without historical data. This study introduces the universal geography neural network, an algorithm designed to glean potential patterns in human mobility across diverse cities and temporal spans. This is achieved through the analysis of substantial quantities of location-based data, resulting in the generation of mobility flows within a city. Our experiment, designed to extract various features and generate fine-grained mobility flows in the testing set, outperforms both traditional models and state-of-the-art deep learning models. Moreover, our model has proven capable of generating reliable results across various time periods and grid areas.","PeriodicalId":156,"journal":{"name":"Computer-Aided Civil and Infrastructure Engineering","volume":"7 1","pages":""},"PeriodicalIF":11.775,"publicationDate":"2025-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142935158","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
For existing bridge weigh-in-motion technologies, the main challenge in accurate weight estimation is to overcome the difficulty of identifying the closely spaced axles. To do so, many field test data are generally required for each bridge in application. To address such a challenge, a novel two-level sequential artificial neural network (ANN) model trained by the hybrid simulated-experimental data was proposed in this study to identify the gross weight and axle weight. For this, simulations and scaled experiments were conducted for the vehicle–bridge interaction system to develop the sequential ANN model. The sequential ANN model was realized by a special data looping strategy, in which the outputs of the global-level ANN served as partial inputs to the following local-level ANN to predict the axle weight. The optimized size of the training data and the appropriate hybrid ratio of the sequential ANN model were also explored. Finally, the proposed algorithm was applied to a real bridge application via transfer learning, as the optimized hybrid sequential ANN model serves as the pre-trained model. The results showed that for the small training datasets with only 5% experimental data, the proposed algorithm significantly improved the accuracy in weight estimation of moving vehicles with closely spaced axles. The field test demonstrated that the proposed algorithm also applies to different bridges within a gross weight identification error of 5%, showing the promise of the proposed algorithm in practical applications.
{"title":"Hybrid-data-driven bridge weigh-in-motion technology using a two-level sequential artificial neural network","authors":"Wangchen Yan, Hao Ren, Xin Luo, Shaofan Li","doi":"10.1111/mice.13415","DOIUrl":"https://doi.org/10.1111/mice.13415","url":null,"abstract":"For existing bridge weigh-in-motion technologies, the main challenge in accurate weight estimation is to overcome the difficulty of identifying the closely spaced axles. To do so, many field test data are generally required for each bridge in application. To address such a challenge, a novel two-level sequential artificial neural network (ANN) model trained by the hybrid simulated-experimental data was proposed in this study to identify the gross weight and axle weight. For this, simulations and scaled experiments were conducted for the vehicle–bridge interaction system to develop the sequential ANN model. The sequential ANN model was realized by a special data looping strategy, in which the outputs of the global-level ANN served as partial inputs to the following local-level ANN to predict the axle weight. The optimized size of the training data and the appropriate hybrid ratio of the sequential ANN model were also explored. Finally, the proposed algorithm was applied to a real bridge application via transfer learning, as the optimized hybrid sequential ANN model serves as the pre-trained model. The results showed that for the small training datasets with only 5% experimental data, the proposed algorithm significantly improved the accuracy in weight estimation of moving vehicles with closely spaced axles. The field test demonstrated that the proposed algorithm also applies to different bridges within a gross weight identification error of 5%, showing the promise of the proposed algorithm in practical applications.","PeriodicalId":156,"journal":{"name":"Computer-Aided Civil and Infrastructure Engineering","volume":"1 1","pages":""},"PeriodicalIF":11.775,"publicationDate":"2025-01-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142929496","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","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：