Measurement最新文献_第7页

In-situ dynamic transmission error -driven lightweight wide-area deconstruction network for gear spalling fault intelligent diagnosis 基于原位动态传动误差驱动的轻量化广域解构网络的齿轮剥落故障智能诊断

IF 5.6 2区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

Measurement

Pub Date : 2026-02-01 DOI: 10.1016/j.measurement.2026.120678

Xuegang Li , Yuanyue Pu , Shengbao Qin , Huayan Pu , Yudong Zhang , Xiaoxi Ding , Wenbin Huang

Gears play a vital role in industrial transmission systems, yet they are prone to degradation and localized faults such as spalling under complex working conditions. Early and accurate fault diagnosis is essential to ensure operational reliability and reduce maintenance costs. Dynamic transmission error (DTE) is defined as the deviation between the actual and theoretical meshing positions of gears. It serves as a key indicator of meshing accuracy and dynamic behavior, directly reflecting the health status of gears. However, conventional analytical approaches for DTE calculation rely on precise physical modeling and known parameters, limiting their adaptability in real-world applications. Meanwhile, encoder-based DTE measurement methods are often impractical for enclosed gearboxes due to installation constraints and susceptibility to environmental interference. To address these limitations, this paper proposes an in-situ DTE-driven lightweight intelligent fault diagnosis (IDLIFD) framework for gear spalling fault diagnosis. First, an in-situ DTE reconstruction and enhancement method is developed to obtain meshing deviation information from in-situ vibration signals, enabling practical DTE measurement in enclosed environments. Then, a lightweight wide-area deconstruction network (WDNet) is designed to extract discriminative spalling-related features from enhanced DTE signals while maintaining a compact structure and low computational complexity. Finally, experimental validation on a self-made gearbox test bench demonstrates that the proposed IDLIFD framework outperforms existing computational methods and lightweight diagnosis models in terms of DTE calculation, diagnosis accuracy, and real-world deployment.

齿轮在工业传动系统中起着至关重要的作用，但在复杂的工作条件下，齿轮容易退化和局部故障，如剥落。早期准确的故障诊断是保证运行可靠性和降低维护成本的关键。动态传动误差（DTE）是指齿轮实际啮合位置与理论啮合位置之间的偏差。它是齿轮啮合精度和动力学行为的关键指标，直接反映齿轮的健康状态。然而，传统的DTE计算分析方法依赖于精确的物理建模和已知的参数，限制了它们在实际应用中的适应性。同时，由于安装限制和易受环境干扰，基于编码器的DTE测量方法通常不适用于封闭式齿轮箱。为了解决这些局限性，本文提出了一种基于原位dte驱动的齿轮脱落故障轻量级智能诊断（IDLIFD）框架。首先，开发了一种原位DTE重建与增强方法，从原位振动信号中获取网格偏差信息，实现了封闭环境下DTE的实际测量。然后，设计了一个轻量级的广域解构网络（WDNet），在保持结构紧凑和低计算复杂度的同时，从增强的DTE信号中提取与碎片相关的判别特征。最后，在自制的齿轮箱试验台上进行了实验验证，结果表明所提出的IDLIFD框架在DTE计算、诊断精度和实际部署方面优于现有的计算方法和轻量级诊断模型。

{"title":"In-situ dynamic transmission error -driven lightweight wide-area deconstruction network for gear spalling fault intelligent diagnosis","authors":"Xuegang Li , Yuanyue Pu , Shengbao Qin , Huayan Pu , Yudong Zhang , Xiaoxi Ding , Wenbin Huang","doi":"10.1016/j.measurement.2026.120678","DOIUrl":"10.1016/j.measurement.2026.120678","url":null,"abstract":"<div><div>Gears play a vital role in industrial transmission systems, yet they are prone to degradation and localized faults such as spalling under complex working conditions. Early and accurate fault diagnosis is essential to ensure operational reliability and reduce maintenance costs. Dynamic transmission error (DTE) is defined as the deviation between the actual and theoretical meshing positions of gears. It serves as a key indicator of meshing accuracy and dynamic behavior, directly reflecting the health status of gears. However, conventional analytical approaches for DTE calculation rely on precise physical modeling and known parameters, limiting their adaptability in real-world applications. Meanwhile, encoder-based DTE measurement methods are often impractical for enclosed gearboxes due to installation constraints and susceptibility to environmental interference. To address these limitations, this paper proposes an in-situ DTE-driven lightweight intelligent fault diagnosis (IDLIFD) framework for gear spalling fault diagnosis. First, an in-situ DTE reconstruction and enhancement method is developed to obtain meshing deviation information from in-situ vibration signals, enabling practical DTE measurement in enclosed environments. Then, a lightweight wide-area deconstruction network (WDNet) is designed to extract discriminative spalling-related features from enhanced DTE signals while maintaining a compact structure and low computational complexity. Finally, experimental validation on a self-made gearbox test bench demonstrates that the proposed IDLIFD framework outperforms existing computational methods and lightweight diagnosis models in terms of DTE calculation, diagnosis accuracy, and real-world deployment.</div></div>","PeriodicalId":18349,"journal":{"name":"Measurement","volume":"268 ","pages":"Article 120678"},"PeriodicalIF":5.6,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146172122","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A robust uncertainty quantification framework for machine learning–based wet-gas flow metering 基于机器学习的湿气流量测量鲁棒不确定性量化框架

IF 5.6 2区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

Measurement

Pub Date : 2026-02-01 DOI: 10.1016/j.measurement.2026.120670

Seyedahmad Hosseini , Gabriele Chinello , Gordon Lindsay , Don McGlinchey

Accurate uncertainty quantification (UQ) is essential for deploying machine learning (ML) models in multiphase flow metering, where limited training data, incomplete feature representation, and distribution shifts across operating conditions introduce significant epistemic uncertainty beyond the inherent variability of the sensors and flow dynamics. Using experimental datasets under diverse multiphase conditions, this study evaluates predictive uncertainty across five ML models, including deep neural networks (DNN), long short-term memory networks (LSTM), random forests (RF), extreme gradient boosting (XGBoost), and Gaussian process regression (GPR). Conformal prediction (CP) is employed as a model-agnostic framework to generate calibrated prediction intervals (PIs), whereas GPR estimates the predictive variance through its kernel-based structure. The evaluation results show that the gas flow rate predictions exhibit high accuracy and well-calibrated intervals across the models, with the CP producing fixed PIs and the GPR achieving the narrowest, dynamically adjusted intervals. Among the CP-based models, RF demonstrated the best balance between the maximum prediction accuracy and minimal uncertainty. Liquid flowrate predictions exhibited improved epistemic uncertainty across all models. In fact, introducing mixture fluid density, a potential engineered feature derived from the gas volume fraction (GVF) and phase densities, decreased uncertainties. The analysis continued using Explainable AI platforms to highlight the importance of features based on predictive strength. The study findings emphasize the importance of both targeted feature engineering and uncertainty-aware modeling, highlighting practical advantages of CP for model-agnostic UQ and the necessity of XAI tools for transparency. Overall, these results support the applicability of explainable, uncertainty-calibrated ML systems for real-time multiphase flow monitoring, with direct implications for metering confidence and operational decision-making.

精确的不确定性量化（UQ）对于在多相流量计量中部署机器学习（ML）模型至关重要，在多相流量计量中，有限的训练数据、不完整的特征表示和跨操作条件的分布变化，除了传感器和流量动力学的固有可变性之外，还会引入显著的认知不确定性。利用不同多相条件下的实验数据集，本研究评估了五种机器学习模型的预测不确定性，包括深度神经网络（DNN）、长短期记忆网络（LSTM）、随机森林（RF）、极端梯度增强（XGBoost）和高斯过程回归（GPR）。共形预测（CP）作为模型不可知的框架来生成校准的预测区间（pi），而探地雷达通过其基于核的结构来估计预测方差。评价结果表明，气体流量预测具有较高的准确性和校准良好的区间，其中CP产生固定的pi，而GPR实现最窄的动态调整区间。在基于cp的模型中，RF在最大预测精度和最小不确定性之间表现出最好的平衡。液体流量预测在所有模型中都表现出更好的认知不确定性。事实上，引入混合流体密度（由气体体积分数（GVF）和相密度衍生的潜在工程特征）可以减少不确定性。分析继续使用Explainable AI平台来突出基于预测强度的特征的重要性。研究结果强调了目标特征工程和不确定性感知建模的重要性，强调了CP对于模型不可知的UQ的实际优势，以及XAI工具对于透明度的必要性。总的来说，这些结果支持了可解释的、不确定度校准的ML系统在实时多相流监测中的适用性，对计量可信度和操作决策有直接影响。

{"title":"A robust uncertainty quantification framework for machine learning–based wet-gas flow metering","authors":"Seyedahmad Hosseini , Gabriele Chinello , Gordon Lindsay , Don McGlinchey","doi":"10.1016/j.measurement.2026.120670","DOIUrl":"10.1016/j.measurement.2026.120670","url":null,"abstract":"<div><div>Accurate uncertainty quantification (UQ) is essential for deploying machine learning (ML) models in multiphase flow metering, where limited training data, incomplete feature representation, and distribution shifts across operating conditions introduce significant epistemic uncertainty beyond the inherent variability of the sensors and flow dynamics. Using experimental datasets under diverse multiphase conditions, this study evaluates predictive uncertainty across five ML models, including deep neural networks (DNN), long short-term memory networks (LSTM), random forests (RF), extreme gradient boosting (XGBoost), and Gaussian process regression (GPR). Conformal prediction (CP) is employed as a model-agnostic framework to generate calibrated prediction intervals (PIs), whereas GPR estimates the predictive variance through its kernel-based structure. The evaluation results show that the gas flow rate predictions exhibit high accuracy and well-calibrated intervals across the models, with the CP producing fixed PIs and the GPR achieving the narrowest, dynamically adjusted intervals. Among the CP-based models, RF demonstrated the best balance between the maximum prediction accuracy and minimal uncertainty. Liquid flowrate predictions exhibited improved epistemic uncertainty across all models. In fact, introducing mixture fluid density, a potential engineered feature derived from the gas volume fraction (GVF) and phase densities, decreased uncertainties. The analysis continued using Explainable AI platforms to highlight the importance of features based on predictive strength. The study findings emphasize the importance of both targeted feature engineering and uncertainty-aware modeling, highlighting practical advantages of CP for model-agnostic UQ and the necessity of XAI tools for transparency. Overall, these results support the applicability of explainable, uncertainty-calibrated ML systems for real-time multiphase flow monitoring, with direct implications for metering confidence and operational decision-making.</div></div>","PeriodicalId":18349,"journal":{"name":"Measurement","volume":"269 ","pages":"Article 120670"},"PeriodicalIF":5.6,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146147329","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Risk-aware multi-response optimization of surface roughness in laser additive manufacturing via conditional value-at-risk desirability modelling 基于条件风险期望值模型的激光增材制造表面粗糙度风险感知多响应优化

IF 5.6 2区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

Measurement

Pub Date : 2026-01-31 DOI: 10.1016/j.measurement.2026.120655

Geetha Narayanan Kannaiyan , Nagasuneetha Darla , Gangadhara Rao Ponugoti , Venkata Phani Babu Vemuri , Bridjesh Pappula , Seshibe Makgato

Surface finish is a decisive quality attribute in direct metal laser sintering, as it governs fatigue life, wear behavior, and the extent of post-processing required. In this study, a Conditional Value-at-Risk Desirability Multi-response Bayesian Optimization (CD-MBO) methodology is applied. This method is designed to penalize parameter settings that produce unstable outcomes and to enable robust multi criteria process decisions by explicitly capturing variability and tail-risk in the surface roughness responses. SS316L specimens were produced according to an L9 orthogonal array on an EOS M 290 system, with laser power, scan speed, and layer thickness considered as the primary process variables. Surface roughness metrics (Ra, Rq, Rz) were subsequently quantified using a Surftest SJ-210 profilometer. The CD-MBO approach aggregated tail-sensitive CVaR_0.80 values of each metric into a weighted desirability function, with uncertainty modeled using a Bayesian bootstrap. The optimal parameter setting was identified as 330 W, 900 mm/s, and 80 µm, yielding Ra = 5.75 ± 0.15 µm, Rq = 6.67 ± 0.01 µm, and Rz = 28.84 ± 0.05 µm, with significantly reduced upper-tail behavior compared to alternative configurations. Sensitivity analyses confirmed that the top-ranked solution was invariant to weighting schemes (variance-based vs. principal component analysis-based) and risk-penalty levels (k = 0 - 0.5). SEM fractography further validated the suppression of porosity and lack-of-fusion defects at the optimal setting. This study makes an innovative contribution to the field through the development of a distribution-aware, metrologically informed optimization methodology for laser powder bed fusion surface finish, which leverages advanced statistical modelling techniques alongside risk-based decision metrics. The study thereby extends the field beyond deterministic quality optimization towards uncertainty-aware measurement decision-making, as it is appropriate to the scope and ambition of the field of Measurement Science.

表面光洁度是直接金属激光烧结的决定性质量属性，因为它决定了疲劳寿命、磨损行为和所需后处理的程度。本研究采用条件风险值可取性多响应贝叶斯优化（CD-MBO）方法。该方法旨在惩罚产生不稳定结果的参数设置，并通过明确捕获表面粗糙度响应的可变性和尾部风险，实现稳健的多标准过程决策。以激光功率、扫描速度和层厚为主要工艺变量，在EOS M 290系统上采用L9正交阵列制备SS316L样品。表面粗糙度指标（Ra, Rq, Rz）随后使用Surftest SJ-210轮廓仪进行量化。CD-MBO方法将每个指标的尾部敏感CVaR0.80值聚合到加权的可取性函数中，并使用贝叶斯自举法对不确定性进行建模。确定最佳参数设置为330 W， 900 mm/s, 80µm，得到Ra = 5.75±0.15µm, Rq = 6.67±0.01µm, Rz = 28.84±0.05µm，与其他配置相比，显著降低了上尾行为。敏感性分析证实，排名靠前的解决方案对加权方案（基于方差的vs.基于主成分分析的）和风险惩罚水平（k = 0 - 0.5）是不变的。SEM断口形貌进一步验证了在最佳设置下孔隙度和未熔合缺陷得到抑制。本研究通过开发一种分布感知、计量信息的激光粉末床融合表面处理优化方法，为该领域做出了创新贡献，该方法利用了先进的统计建模技术和基于风险的决策指标。因此，该研究将该领域从确定性质量优化扩展到不确定性感知的测量决策，因为它适合测量科学领域的范围和抱负。

{"title":"Risk-aware multi-response optimization of surface roughness in laser additive manufacturing via conditional value-at-risk desirability modelling","authors":"Geetha Narayanan Kannaiyan , Nagasuneetha Darla , Gangadhara Rao Ponugoti , Venkata Phani Babu Vemuri , Bridjesh Pappula , Seshibe Makgato","doi":"10.1016/j.measurement.2026.120655","DOIUrl":"10.1016/j.measurement.2026.120655","url":null,"abstract":"<div><div>Surface finish is a decisive quality attribute in direct metal laser sintering, as it governs fatigue life, wear behavior, and the extent of post-processing required. In this study, a Conditional Value-at-Risk Desirability Multi-response Bayesian Optimization (CD-MBO) methodology is applied. This method is designed to penalize parameter settings that produce unstable outcomes and to enable robust multi criteria process decisions by explicitly capturing variability and tail-risk in the surface roughness responses. SS316L specimens were produced according to an L9 orthogonal array on an EOS M 290 system, with laser power, scan speed, and layer thickness considered as the primary process variables. Surface roughness metrics (Ra, Rq, Rz) were subsequently quantified using a Surftest SJ-210 profilometer. The CD-MBO approach aggregated tail-sensitive CVaR<sub>0.80</sub> values of each metric into a weighted desirability function, with uncertainty modeled using a Bayesian bootstrap. The optimal parameter setting was identified as 330 W, 900 mm/s, and 80 µm, yielding Ra = 5.75 ± 0.15 µm, Rq = 6.67 ± 0.01 µm, and Rz = 28.84 ± 0.05 µm, with significantly reduced upper-tail behavior compared to alternative configurations. Sensitivity analyses confirmed that the top-ranked solution was invariant to weighting schemes (variance-based vs. principal component analysis-based) and risk-penalty levels (<em>k</em> = 0 - 0.5). SEM fractography further validated the suppression of porosity and lack-of-fusion defects at the optimal setting. This study makes an innovative contribution to the field through the development of a distribution-aware, metrologically informed optimization methodology for laser powder bed fusion surface finish, which leverages advanced statistical modelling techniques alongside risk-based decision metrics. The study thereby extends the field beyond deterministic quality optimization towards uncertainty-aware measurement decision-making, as it is appropriate to the scope and ambition of the field of Measurement Science.</div></div>","PeriodicalId":18349,"journal":{"name":"Measurement","volume":"268 ","pages":"Article 120655"},"PeriodicalIF":5.6,"publicationDate":"2026-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146172029","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Electromagnetic noise characterization in MRI-guided microwave ablation 磁共振引导微波消融术中的电磁噪声表征

IF 5.6 2区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

Measurement

Pub Date : 2026-01-31 DOI: 10.1016/j.measurement.2026.120607

Daniel Hernandez , Taewoo Nam , Eunwoo Lee , Aiming Lu , Christopher P. Favazza , Eric G. Stinson , David A. Woodrum , Myung-Ho In , Kyoung-Nam Kim

Microwave ablation (MWA) guided by magnetic resonance imaging (MRI) is an effective approach for minimally invasive tumor treatment, combining MRI’s soft tissue image contrast and temperature mapping capabilities with the ablative power of microwave energy. However, MRI-guided MWA faces a significant challenge with image noise generated by electromagnetic (EM) interference, which degrades image quality and limits real-time monitoring accuracy. As an initial step to address this problem, we introduce a novel noise characterization method for MRI-guided MWA using EM modeling and simulation. Empirically acquired noise data from controlled phantom experiments was used to develop a theoretical framework for simulating interactions between MRI components and the MWA device. The simulations included key system components: the RF transmit coil, receiver coil, phantom, and MWA probe, and produced a signal–noise map that closely matched the experimental data, effectively replicating observed noise patterns. To demonstrate the use of these simulations for practical applications, we evaluated a simple filter circuit for its effectiveness in reducing noise and validated simulation results through benchwork, which showed significant improvements. The results suggest that this approach provides valuable insights into the underlying noise mechanisms and can inform potential strategies for noise mitigation, offering a practical tool for optimizing MRI-guided MWA and enhancing the efficacy of interventional MRI procedures.

磁共振成像（MRI）引导下的微波消融（MWA）是一种有效的微创肿瘤治疗方法，它将MRI的软组织图像对比和温度成像能力与微波能量的消融能力相结合。然而，mri引导的MWA面临着由电磁干扰（EM）产生的图像噪声的重大挑战，这些噪声会降低图像质量并限制实时监测精度。作为解决这一问题的第一步，我们引入了一种利用EM建模和仿真的新型mri引导MWA噪声表征方法。从受控幻象实验中获得的经验噪声数据用于建立模拟MRI组件与MWA设备之间相互作用的理论框架。模拟包括关键的系统组件：射频发射线圈、接收线圈、模体和MWA探针，并生成了与实验数据密切匹配的信噪图，有效地复制了观察到的噪声模式。为了证明这些模拟在实际应用中的应用，我们评估了一个简单的滤波电路在降低噪声方面的有效性，并通过基准测试验证了模拟结果，结果显示出显着的改进。结果表明，该方法为潜在的噪声机制提供了有价值的见解，可以为潜在的噪声缓解策略提供信息，为优化MRI引导下的MWA和提高介入性MRI手术的有效性提供了实用工具。

{"title":"Electromagnetic noise characterization in MRI-guided microwave ablation","authors":"Daniel Hernandez , Taewoo Nam , Eunwoo Lee , Aiming Lu , Christopher P. Favazza , Eric G. Stinson , David A. Woodrum , Myung-Ho In , Kyoung-Nam Kim","doi":"10.1016/j.measurement.2026.120607","DOIUrl":"10.1016/j.measurement.2026.120607","url":null,"abstract":"<div><div>Microwave ablation (MWA) guided by magnetic resonance imaging (MRI) is an effective approach for minimally invasive tumor treatment, combining MRI’s soft tissue image contrast and temperature mapping capabilities with the ablative power of microwave energy. However, MRI-guided MWA faces a significant challenge with image noise generated by electromagnetic (EM) interference, which degrades image quality and limits real-time monitoring accuracy. As an initial step to address this problem, we introduce a novel noise characterization method for MRI-guided MWA using EM modeling and simulation. Empirically acquired noise data from controlled phantom experiments was used to develop a theoretical framework for simulating interactions between MRI components and the MWA device. The simulations included key system components: the RF transmit coil, receiver coil, phantom, and MWA probe, and produced a signal–noise map that closely matched the experimental data, effectively replicating observed noise patterns. To demonstrate the use of these simulations for practical applications, we evaluated a simple filter circuit for its effectiveness in reducing noise and validated simulation results through benchwork, which showed significant improvements. The results suggest that this approach provides valuable insights into the underlying noise mechanisms and can inform potential strategies for noise mitigation, offering a practical tool for optimizing MRI-guided MWA and enhancing the efficacy of interventional MRI procedures.</div></div>","PeriodicalId":18349,"journal":{"name":"Measurement","volume":"268 ","pages":"Article 120607"},"PeriodicalIF":5.6,"publicationDate":"2026-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146096171","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Research on real-time monocular visual positioning system for mobile robotic construction with automated error compensation 基于误差自动补偿的移动施工机器人实时单目视觉定位系统研究

IF 5.6 2区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

Measurement

Pub Date : 2026-01-31 DOI: 10.1016/j.measurement.2026.120643

Kang Bi , Xinyu Shi , Da Wan , Weijiu Cui , Haining Zhou , Chengpeng Sun , Peng Du , Hiroatsu Fukuda

Accurate end-effector positioning is crucial for mobile robotic platforms (MCP) in automated on-site construction tasks. This paper introduces a real-time visual positioning system based on the AprilTag algorithm, specifically developed to enhance the end-effector accuracy of MCPs used in construction. The solution builds upon an autonomous navigation MCP platform by incorporating Fiducial Marker System (FMS) and automatic interpolation compensation algorithm based on camera-specific error model. Experimental results demonstrate a significant improvement in accuracy: the enhanced algorithm achieved an average positioning accuracy of 1.08 mm—17.56% improvement over the native AprilTag method—and positioning stability improved by 85.7% (with a standard deviation of 0.38 mm). Comprehensive experiments, including random point positioning, curve fitting, and physical assembly tasks, confirmed the system’s robustness, repeatability, and industrial applicability. This method enables MCPs to autonomously adjust robot working path in real time according to dynamic on-site conditions, adapt to unpredictable construction environments, and significantly enhance construction precision.

准确的末端执行器定位是移动机器人平台（MCP）在自动化现场施工任务中的关键。本文介绍了一种基于AprilTag算法的实时视觉定位系统，该系统是专门为提高建筑用mcp末端执行器的精度而开发的。该解决方案建立在自主导航MCP平台上，结合了基准标记系统（FMS）和基于相机特定误差模型的自动插值补偿算法。实验结果表明，改进后的算法在精度上有了显著的提高，平均定位精度比原生AprilTag方法提高了1.08 mm - 17.56%，定位稳定性提高了85.7%（标准差为0.38 mm）。综合实验，包括随机点定位、曲线拟合和物理装配任务，证实了系统的鲁棒性、可重复性和工业适用性。该方法使mcp能够根据现场动态情况实时自主调整机器人工作路径，适应不可预测的施工环境，显著提高施工精度。

{"title":"Research on real-time monocular visual positioning system for mobile robotic construction with automated error compensation","authors":"Kang Bi , Xinyu Shi , Da Wan , Weijiu Cui , Haining Zhou , Chengpeng Sun , Peng Du , Hiroatsu Fukuda","doi":"10.1016/j.measurement.2026.120643","DOIUrl":"10.1016/j.measurement.2026.120643","url":null,"abstract":"<div><div>Accurate end-effector positioning is crucial for mobile robotic platforms (MCP) in automated on-site construction tasks. This paper introduces a real-time visual positioning system based on the AprilTag algorithm, specifically developed to enhance the end-effector accuracy of MCPs used in construction. The solution builds upon an autonomous navigation MCP platform by incorporating Fiducial Marker System (FMS) and automatic interpolation compensation algorithm based on camera-specific error model. Experimental results demonstrate a significant improvement in accuracy: the enhanced algorithm achieved an average positioning accuracy of 1.08 mm—17.56% improvement over the native AprilTag method—and positioning stability improved by 85.7% (with a standard deviation of 0.38 mm). Comprehensive experiments, including random point positioning, curve fitting, and physical assembly tasks, confirmed the system’s robustness, repeatability, and industrial applicability. This method enables MCPs to autonomously adjust robot working path in real time according to dynamic on-site conditions, adapt to unpredictable construction environments, and significantly enhance construction precision.</div></div>","PeriodicalId":18349,"journal":{"name":"Measurement","volume":"268 ","pages":"Article 120643"},"PeriodicalIF":5.6,"publicationDate":"2026-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146172035","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Application of an improved diffusion model-based data augmentation method in partial discharge pattern recognition for vehicle cable terminals 一种改进的基于扩散模型的数据增强方法在车载电缆终端局部放电模式识别中的应用

IF 5.6 2区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

Measurement

Pub Date : 2026-01-31 DOI: 10.1016/j.measurement.2026.120672

Tingyu Zhang, Kai Liu, Guangbo Nie, Guoqiang Gao, Guangning Wu

Difficult data collection and limited labeled samples of partial discharge (PD) signals from vehicle cable terminals lead to insufficient recognition accuracy and poor generalization capability in pattern recognition networks. To address these issues, this study proposes a data augmentation method based on image generation. Initially, the Markov Transition Field (MTF) serves to translate the original one-dimensional PD signals into two-dimensional images, building up a foundational PD sample library. Furthermore, within the diffusion model framework, conditional information and supervised contrastive learning mechanisms are innovatively integrated to achieve high-quality conditional PD sample generation. Experiments demonstrate that the PD images generated by this method are of significantly higher quality than those generated by other comparative models. Augmenting the original sample library with the generated samples effectively improves the recognition performance of various classification models. Finally, an accuracy of up to 98.54% in PD pattern classification is achieved using an improved residual network model, significantly enhancing the PD diagnosis capability for vehicle cable terminals.

车辆电缆终端局部放电信号的数据采集困难，标记样本有限，导致模式识别网络的识别精度不高，泛化能力差。为了解决这些问题，本研究提出了一种基于图像生成的数据增强方法。最初，马尔可夫过渡场（MTF）用于将原始一维PD信号转换为二维图像，建立基础PD样本库。此外，在扩散模型框架内，创新地集成了条件信息和监督对比学习机制，以实现高质量的条件PD样本生成。实验表明，该方法生成的PD图像质量明显高于其他比较模型生成的PD图像质量。用生成的样本对原始样本库进行扩充，有效地提高了各种分类模型的识别性能。最后，利用改进的残差网络模型对PD模式进行分类，准确率高达98.54%，显著提高了车载电缆终端PD诊断能力。

{"title":"Application of an improved diffusion model-based data augmentation method in partial discharge pattern recognition for vehicle cable terminals","authors":"Tingyu Zhang, Kai Liu, Guangbo Nie, Guoqiang Gao, Guangning Wu","doi":"10.1016/j.measurement.2026.120672","DOIUrl":"10.1016/j.measurement.2026.120672","url":null,"abstract":"<div><div>Difficult data collection and limited labeled samples of partial discharge (PD) signals from vehicle cable terminals lead to insufficient recognition accuracy and poor generalization capability in pattern recognition networks. To address these issues, this study proposes a data augmentation method based on image generation. Initially, the Markov Transition Field (MTF) serves to translate the original one-dimensional PD signals into two-dimensional images, building up a foundational PD sample library. Furthermore, within the diffusion model framework, conditional information and supervised contrastive learning mechanisms are innovatively integrated to achieve high-quality conditional PD sample generation. Experiments demonstrate that the PD images generated by this method are of significantly higher quality than those generated by other comparative models. Augmenting the original sample library with the generated samples effectively improves the recognition performance of various classification models. Finally, an accuracy of up to 98.54% in PD pattern classification is achieved using an improved residual network model, significantly enhancing the PD diagnosis capability for vehicle cable terminals.</div></div>","PeriodicalId":18349,"journal":{"name":"Measurement","volume":"268 ","pages":"Article 120672"},"PeriodicalIF":5.6,"publicationDate":"2026-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146172139","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Experimental study on three-dimensional deformation reconstruction of pipelines using the inverse finite element method 基于逆有限元法的管道三维变形重建试验研究

IF 5.6 2区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

Measurement

Pub Date : 2026-01-31 DOI: 10.1016/j.measurement.2026.120659

Yuanju Wang, Runzhou You, Liang Ren, Xin Feng, Jiang Cui

Pipeline deformation monitoring plays a crucial role in ensuring the safety of oil and gas transportation. The inverse finite element method (iFEM) is an emerging deformation reconstruction approach that can be implemented without invoking force equilibrium. While limited experimental work is available for three-dimensional deformation reconstruction of pipelines. To overcome this limitation, this paper proposes an integrated approach combining the inverse finite element method (iFEM) with FBG/DOFS-based strain sensing technology for three-dimensional deformation reconstruction of pipelines. Two experimental studies were conducted, encompassing the three-dimensional deformation evaluative testing and the large-scale model testing. During the three-dimensional deformation evaluative testing, a numerical model was established, and the effect of strain sensor location and discrete element level was examined versus the solution accuracy. Next, the optimized strain sensor position derived from simulation analysis was applied to instruct experiments for the three-dimensional deformation monitoring problems under the concentrated loads, and the effectiveness of the method was demonstrated by comparing the reconstructed results with those of dial indicator set. Additionally, the large-scale model testing system was experimentally applied to monitor and analyze the deformation behavior of pipeline structures crossing a strike-slip fault. Subsequently, these dense strain measurements from distributed optical fiber sensors were used to perform iFEM analysis, which utilized various discrete elements and strain sensor configurations. Overall, experimental results demonstrate that this method facilitates localization of the three-dimensional pipeline deformations, aiding pipeline condition assessment and exhibiting the potential for practical applications.

管道变形监测对保障油气运输安全具有至关重要的作用。逆有限元法（iFEM）是一种新兴的变形重建方法，可以在不调用力平衡的情况下实现。而管道三维变形重建的实验工作有限。为了克服这一局限性，本文提出了一种将反有限元法（iFEM）与基于FBG/ dofs的应变传感技术相结合的管道三维变形重建方法。进行了三维变形评价试验和大尺度模型试验两项试验研究。在三维变形评估试验中，建立了数值模型，考察了应变传感器位置和离散单元水平对求解精度的影响。然后，将仿真分析得到的优化应变传感器位置用于集中载荷下三维变形监测问题的实验指导，并将重建结果与刻度盘组结果进行对比，验证了该方法的有效性。在此基础上，利用大尺度模型试验系统对管道结构穿越走滑断层的变形行为进行了实验监测和分析。随后，利用这些来自分布式光纤传感器的密集应变测量数据进行iFEM分析，该分析使用了各种离散单元和应变传感器配置。实验结果表明，该方法有助于管道三维变形的定位，有助于管道状态评估，具有实际应用潜力。

{"title":"Experimental study on three-dimensional deformation reconstruction of pipelines using the inverse finite element method","authors":"Yuanju Wang, Runzhou You, Liang Ren, Xin Feng, Jiang Cui","doi":"10.1016/j.measurement.2026.120659","DOIUrl":"10.1016/j.measurement.2026.120659","url":null,"abstract":"<div><div>Pipeline deformation monitoring plays a crucial role in ensuring the safety of oil and gas transportation. The inverse finite element method (iFEM) is an emerging deformation reconstruction approach that can be implemented without invoking force equilibrium. While limited experimental work is available for three-dimensional deformation reconstruction of pipelines. To overcome this limitation, this paper proposes an integrated approach combining the inverse finite element method (iFEM) with FBG/DOFS-based strain sensing technology for three-dimensional deformation reconstruction of pipelines. Two experimental studies were conducted, encompassing the three-dimensional deformation evaluative testing and the large-scale model testing. During the three-dimensional deformation evaluative testing, a numerical model was established, and the effect of strain sensor location and discrete element level was examined versus the solution accuracy. Next, the optimized strain sensor position derived from simulation analysis was applied to instruct experiments for the three-dimensional deformation monitoring problems under the concentrated loads, and the effectiveness of the method was demonstrated by comparing the reconstructed results with those of dial indicator set. Additionally, the large-scale model testing system was experimentally applied to monitor and analyze the deformation behavior of pipeline structures crossing a strike-slip fault. Subsequently, these dense strain measurements from distributed optical fiber sensors were used to perform iFEM analysis, which utilized various discrete elements and strain sensor configurations. Overall, experimental results demonstrate that this method facilitates localization of the three-dimensional pipeline deformations, aiding pipeline condition assessment and exhibiting the potential for practical applications.</div></div>","PeriodicalId":18349,"journal":{"name":"Measurement","volume":"268 ","pages":"Article 120659"},"PeriodicalIF":5.6,"publicationDate":"2026-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146172090","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Improved time-correlated noise modeling for GNSS terrestrial reference frame realization via square root information filter 基于平方根信息滤波的GNSS地面参考帧改进时相关噪声建模

IF 5.6 2区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

Measurement

Pub Date : 2026-01-31 DOI: 10.1016/j.measurement.2026.120644

Yanlin Li , Na Wei , Guo Chen , Chuang Shi , Jingnan Liu

Realizing a Terrestrial Reference Frame (TRF) with an accuracy of 1 mm and a long-term stability of 0.1 mm/yr is a longstanding goal of the geodesy field. To achieve this, selecting an appropriate stochastic model to accurately characterize the nonlinear coordinate variations of geodetic stations is essential for TRF realization. However, the commonly used Random Walk (RW) model in filtering is not the optimal noise model for time-correlated noise in Global Navigation Satellite System (GNSS) coordinates. In this study, we replace the RW model with a first-order autoregressive (AR[1]) process to model the GNSS time-correlated noise and implement a GNSS TRF solution aligned with ITRF2020 via the Square Root Information Filter (SRIF). We found that the AR[1] process used in this study has a higher cut-off frequency than the RW model, allowing it to retain a larger portion of the input flicker noise. Consequently, the GNSS time-correlated noise modelled by AR[1] more closely approximates true flicker noise than that modelled by RW. When time-correlated noise is modelled by AR[1], the median RMS of coordinate residuals is decreases to 0.3 and 2.0 mm in the horizontal and up components, respectively. Moreover, the AR[1] process can capture short-term correlations in time-correlated noise parameters, thereby enhancing the accuracy of short-term (approximately 11 weeks) TRF coordinate predictions. These findings demonstrate the potential of incorporating time-correlated noise using AR[1] in GNSS data assimilation, with implications for both multi-technique global TRF realization and regional GNSS TRF solutions.

实现精度为1毫米、长期稳定性为0.1毫米/年的地面参考系（TRF）是大地测量学领域的长期目标。为了实现这一目标，选择合适的随机模型来准确表征测地站的非线性坐标变化是TRF实现的关键。然而，对于全球导航卫星系统（GNSS）坐标下的时间相关噪声，常用的随机漫步（Random Walk， RW）滤波模型并不是最优的噪声模型。在本研究中，我们用一阶自回归（AR[1]）过程代替RW模型来模拟GNSS时间相关噪声，并通过平方根信息滤波器（SRIF）实现与ITRF2020一致的GNSS TRF解决方案。我们发现，本研究中使用的AR[1]过程比RW模型具有更高的截止频率，使其能够保留更大一部分输入闪烁噪声。因此，与RW模型相比，AR[1]模型的GNSS时间相关噪声更接近真实闪烁噪声。当时间相关噪声用AR[1]建模时，坐标残差的中位数均方根值在水平分量和向上分量分别降至0.3和2.0 mm。此外，AR[1]过程可以捕获时间相关噪声参数的短期相关性，从而提高短期（约11周）后机匣坐标预测的准确性。这些发现证明了在GNSS数据同化中使用AR[1]结合时间相关噪声的潜力，这对多技术全球TRF实现和区域GNSS TRF解决方案都具有重要意义。

{"title":"Improved time-correlated noise modeling for GNSS terrestrial reference frame realization via square root information filter","authors":"Yanlin Li , Na Wei , Guo Chen , Chuang Shi , Jingnan Liu","doi":"10.1016/j.measurement.2026.120644","DOIUrl":"10.1016/j.measurement.2026.120644","url":null,"abstract":"<div><div>Realizing a Terrestrial Reference Frame (TRF) with an accuracy of 1 mm and a long-term stability of 0.1 mm/yr is a longstanding goal of the geodesy field. To achieve this, selecting an appropriate stochastic model to accurately characterize the nonlinear coordinate variations of geodetic stations is essential for TRF realization. However, the commonly used Random Walk (RW) model in filtering is not the optimal noise model for time-correlated noise in Global Navigation Satellite System (GNSS) coordinates. In this study, we replace the RW model with a first-order autoregressive (AR[1]) process to model the GNSS time-correlated noise and implement a GNSS TRF solution aligned with ITRF2020 via the Square Root Information Filter (SRIF). We found that the AR[1] process used in this study has a higher cut-off frequency than the RW model, allowing it to retain a larger portion of the input flicker noise. Consequently, the GNSS time-correlated noise modelled by AR[1] more closely approximates true flicker noise than that modelled by RW. When time-correlated noise is modelled by AR[1], the median RMS of coordinate residuals is decreases to 0.3 and 2.0 mm in the horizontal and up components, respectively. Moreover, the AR[1] process can capture short-term correlations in time-correlated noise parameters, thereby enhancing the accuracy of short-term (approximately 11 weeks) TRF coordinate predictions. These findings demonstrate the potential of incorporating time-correlated noise using AR[1] in GNSS data assimilation, with implications for both multi-technique global TRF realization and regional GNSS TRF solutions.</div></div>","PeriodicalId":18349,"journal":{"name":"Measurement","volume":"268 ","pages":"Article 120644"},"PeriodicalIF":5.6,"publicationDate":"2026-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146172086","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

The contribution of satellite laser ranging to the BDS-3 constellation: precise orbit determination and geodetic parameters estimation 卫星激光测距对北斗三号星座的贡献：精确定轨和大地参数估计

IF 5.6 2区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

Measurement

Pub Date : 2026-01-31 DOI: 10.1016/j.measurement.2026.120645

Chen Ren , Chen Wang , Zhenhong Li , Mingrui Yang , Haoran Gong

While satellite laser ranging (SLR) has long been recognized as a valuable complement to microwave tracking for GNSS precise orbit determination, previous studies have largely examined individual modeling aspects in isolation. A systematic investigation of how SLR prior precision, range-bias (RB) parameterization, and their prior constraints jointly affect BDS-3 orbit determination and geodetic parameters is still lacking. Using four years of BDS-3 microwave and SLR observations, this study systematically assesses the impact of different SLR–BDS-3 fusion strategies on precise orbit determination, geocenter motion, and Earth rotation parameters (ERPs). We identify an reliable strategy (denoted as M1C5) in which station–satellite RB are estimated together with prior sigmas of 1 cm for SLR observations and 5 cm for RB. This strategy yields stable and significant improvements in orbit accuracy and geocenter motion compared to the microwave-only solution. The contribution of SLR to radial orbit accuracy is strongly anti-correlated with the density of the microwave tracking network; when the global network expands to 40 stations, the additional SLR-induced improvement is reduced to about 0.59%. Spectral analysis of the geocenter time series further indicates that inappropriate RB modeling amplifies spurious short-period noise, including an artificial 7-day signal and other short-period oscillations, whereas the M1C5 strategy effectively suppresses these artifacts. Moreover, incorporating SLR enhances the recovery of annual and sub-annual geocenter harmonics in the X, Y, and Z components. In contrast, the impact of SLR on ERP estimation is marginal and not systematic under the current data volume and network geometry.

虽然卫星激光测距（SLR）长期以来一直被认为是GNSS精确定轨微波跟踪的宝贵补充，但以前的研究在很大程度上是孤立地考察了单个建模方面。对于单反先验精度、距离偏差（RB）参数化及其先验约束如何共同影响北斗三号系统定轨和大地测量参数，尚缺乏系统的研究。利用4年的BDS-3微波和单反卫星观测数据，系统评估了不同的SLR - BDS-3融合策略对精确定轨、中心运动和地球自转参数（ERPs）的影响。我们确定了一种可靠的策略（表示为M1C5），其中站星RB与SLR观测的1厘米和RB的5厘米的先验sigma一起估计。与仅使用微波的解决方案相比，这种策略在轨道精度和地心运动方面产生了稳定而显著的改进。单反对径向轨道精度的贡献与微波跟踪网络密度呈强反相关；当全球网络扩展到40个站点时，slr引起的额外改善减少到0.59%左右。对地球中心时间序列的频谱分析进一步表明，不适当的RB建模放大了虚假的短周期噪声，包括人为的7天信号和其他短周期振荡，而M1C5策略有效地抑制了这些伪影。此外，加入单反可以增强X、Y和Z分量的年和次年地心谐波的恢复。相比之下，在当前的数据量和网络几何结构下，SLR对ERP估计的影响是边际的，不系统的。

{"title":"The contribution of satellite laser ranging to the BDS-3 constellation: precise orbit determination and geodetic parameters estimation","authors":"Chen Ren , Chen Wang , Zhenhong Li , Mingrui Yang , Haoran Gong","doi":"10.1016/j.measurement.2026.120645","DOIUrl":"10.1016/j.measurement.2026.120645","url":null,"abstract":"<div><div>While satellite laser ranging (SLR) has long been recognized as a valuable complement to microwave tracking for GNSS precise orbit determination, previous studies have largely examined individual modeling aspects in isolation. A systematic investigation of how SLR prior precision, range-bias (RB) parameterization, and their prior constraints jointly affect BDS-3 orbit determination and geodetic parameters is still lacking. Using four years of BDS-3 microwave and SLR observations, this study systematically assesses the impact of different SLR–BDS-3 fusion strategies on precise orbit determination, geocenter motion, and Earth rotation parameters (ERPs). We identify an reliable strategy (denoted as M1C5) in which station–satellite RB are estimated together with prior sigmas of 1 cm for SLR observations and 5 cm for RB. This strategy yields stable and significant improvements in orbit accuracy and geocenter motion compared to the microwave-only solution. The contribution of SLR to radial orbit accuracy is strongly anti-correlated with the density of the microwave tracking network; when the global network expands to 40 stations, the additional SLR-induced improvement is reduced to about 0.59%. Spectral analysis of the geocenter time series further indicates that inappropriate RB modeling amplifies spurious short-period noise, including an artificial 7-day signal and other short-period oscillations, whereas the M1C5 strategy effectively suppresses these artifacts. Moreover, incorporating SLR enhances the recovery of annual and sub-annual geocenter harmonics in the X, Y, and Z components. In contrast, the impact of SLR on ERP estimation is marginal and not systematic under the current data volume and network geometry.</div></div>","PeriodicalId":18349,"journal":{"name":"Measurement","volume":"268 ","pages":"Article 120645"},"PeriodicalIF":5.6,"publicationDate":"2026-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146172040","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Missile remaining flight time estimation via physics-guided residual learning 基于物理制导残差学习的导弹剩余飞行时间估计

IF 5.6 2区工程技术 Q1 ENGINEERING, MULTIDISCIPLINARY

Measurement

Pub Date : 2026-01-31 DOI: 10.1016/j.measurement.2026.120677

Zhanpeng Gao , Jun Liu , Wenjun Yi , Shusen Yuan , Jun Guan

Addressing the challenge of insufficient accuracy in time-to-go (

t_{g o}

) estimation for missiles in complex combat environments is critical. While existing physical model-based methods are computationally efficient, they often struggle to effectively cope with high-dynamic and highly uncertain battlefield conditions. To overcome this limitation, this paper proposes a hybrid error compensation model that employs the Harris Hawks Optimization (HHO) algorithm to optimize the Extreme Gradient Boosting (XGBoost) framework. This method utilizes HHO-XGBoost to learn the nonlinear deviation between the physical analytical solution and the actual flight time, providing real-time compensation to the physical model. Simulation results demonstrate that the proposed model exhibits superior accuracy: in stationary target engagement missions, the maximum prediction error is merely 0.45 s; in more complex moving target scenarios, the maximum error is controlled within 0.27 s, significantly outperforming existing comparative models. Furthermore, the model maintains high prediction stability under varying degrees of observation errors, verifying its strong robustness. Application of this algorithm to an Impact Time Control guidance law reveals that high-precision

t_{g o}

estimation not only ensures precise target impact at the desired time but also effectively avoids acceleration saturation caused by estimation deviations during the initial guidance phase, thereby significantly enhancing the overall stability of the guidance process

解决复杂作战环境下导弹预估tgo精度不足的问题至关重要。虽然现有的基于物理模型的方法计算效率很高，但它们往往难以有效地应对高动态和高度不确定的战场条件。为了克服这一限制，本文提出了一种混合误差补偿模型，该模型采用Harris Hawks Optimization （HHO）算法来优化Extreme Gradient Boosting （XGBoost）框架。该方法利用HHO-XGBoost学习物理解析解与实际飞行时间之间的非线性偏差，对物理模型进行实时补偿。仿真结果表明，该模型具有较高的精度：在静止目标交战任务中，最大预测误差仅为0.45 s；在更复杂的运动目标场景下，最大误差控制在0.27 s以内，明显优于现有的比较模型。此外，该模型在不同程度的观测误差下仍保持较高的预测稳定性，验证了其较强的鲁棒性。该算法在冲击时间控制制导律中的应用表明，高精度tgo估计不仅保证了目标在期望时间的精确撞击，而且有效避免了初始制导阶段估计偏差导致的加速度饱和，从而显著提高了制导过程的整体稳定性

{"title":"Missile remaining flight time estimation via physics-guided residual learning","authors":"Zhanpeng Gao , Jun Liu , Wenjun Yi , Shusen Yuan , Jun Guan","doi":"10.1016/j.measurement.2026.120677","DOIUrl":"10.1016/j.measurement.2026.120677","url":null,"abstract":"<div><div>Addressing the challenge of insufficient accuracy in time-to-go (<span><math><msub><mrow><mi>t</mi></mrow><mrow><mi>g</mi><mi>o</mi></mrow></msub></math></span>) estimation for missiles in complex combat environments is critical. While existing physical model-based methods are computationally efficient, they often struggle to effectively cope with high-dynamic and highly uncertain battlefield conditions. To overcome this limitation, this paper proposes a hybrid error compensation model that employs the Harris Hawks Optimization (HHO) algorithm to optimize the Extreme Gradient Boosting (XGBoost) framework. This method utilizes HHO-XGBoost to learn the nonlinear deviation between the physical analytical solution and the actual flight time, providing real-time compensation to the physical model. Simulation results demonstrate that the proposed model exhibits superior accuracy: in stationary target engagement missions, the maximum prediction error is merely 0.45 s; in more complex moving target scenarios, the maximum error is controlled within 0.27 s, significantly outperforming existing comparative models. Furthermore, the model maintains high prediction stability under varying degrees of observation errors, verifying its strong robustness. Application of this algorithm to an Impact Time Control guidance law reveals that high-precision <span><math><msub><mrow><mi>t</mi></mrow><mrow><mi>g</mi><mi>o</mi></mrow></msub></math></span> estimation not only ensures precise target impact at the desired time but also effectively avoids acceleration saturation caused by estimation deviations during the initial guidance phase, thereby significantly enhancing the overall stability of the guidance process</div></div>","PeriodicalId":18349,"journal":{"name":"Measurement","volume":"268 ","pages":"Article 120677"},"PeriodicalIF":5.6,"publicationDate":"2026-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146172134","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0