The multi-point dynamic aggregation(MPDA)problem is a challenging real-world problem.In the MPDA problem,the demands of tasks keep changing with their inherent incremental rates,while a heterogeneous robot fleet is re...The multi-point dynamic aggregation(MPDA)problem is a challenging real-world problem.In the MPDA problem,the demands of tasks keep changing with their inherent incremental rates,while a heterogeneous robot fleet is required to travel between these tasks to change the time-varying state of each task.The robots are allowed to collaborate on the same task or work separately until all tasks are completed.It is challenging to generate an effective task execution plan due to the tight coupling between robots abilities and tasks'incremental rates,and the complexity of robot collaboration.For effectiveness consideration,we use the variable length encoding to avoid redundancy in the solution space.We creatively use the adaptive large neighborhood search(ALNS)framework to solve the MPDA problem.In the proposed algorithm,high-quality initial solutions are generated through multiple problem-specific solution construction heuristics.These heuristics are also used to fix the broken solution in the novel integrated decoding-construction repair process of the ALNS framework.The results of statistical analysis by the Wilcoxon rank-sum test demonstrate that the proposed ALNS can obtain better task execution plans than some state-of-the-art algorithms in most MPDA instances.展开更多
Sandwich structures are vulnerable to multi-point impacts,and such impacts can result in a reduction in residual strength even catastrophic accident.Therefore,the multi-point impact behaviors of PMI foam sandwich stru...Sandwich structures are vulnerable to multi-point impacts,and such impacts can result in a reduction in residual strength even catastrophic accident.Therefore,the multi-point impact behaviors of PMI foam sandwich structure are investigated and studied using experimental and numerical coupled methods.Three impact energy levels and five Distances Between Impact Positions(DBIP)are considered in details,and representative impact characteristics are compared to reveal the association between Compression After Impact(CAI)strength and DBIP.Results indicate that the interference between the multi-point impact events has a dominant effect on CAI strength when DBIP is small,and the variation in bending stiffness induced by the boundary effect is the dominant factor affecting CAI strength when DBIP ranges from 20 mm to 60 mm.In addition,matrix damage represents the primary damage mode in multi-point impact,and the calculated ratio of energy absorbed by the top face sheet and honeycomb core,in relation to the total absorbed energy,serves as a clear indicator of the damage severity experienced by both components.This work is enlightening for the structural design of impact-resistant composites.展开更多
Purpose – Straightness measurement of rail weld joint is of essential importance to railway maintenance. Dueto the lack of efficient measurement equipment, there has been limited in-depth research on rail weld joint ...Purpose – Straightness measurement of rail weld joint is of essential importance to railway maintenance. Dueto the lack of efficient measurement equipment, there has been limited in-depth research on rail weld joint with a5-m wavelength range, leaving a significant knowledge gap in this field.Design/methodology/approach – In this study, the authors used the well-established inertial referencemethod (IR-method), and the state-of-the-art multi-point chord reference method (MCR-method). Two methodshave been applied in different types of rail straightness measurement trollies, respectively. These instrumentswere tested in a high-speed rail section within a certain region of China. The test results were ultimatelyvalidated through using traditional straightedge and feeler gauge methods as reference data to evaluate the railweld joint straightness within the 5-m wavelength range.Findings – The research reveals that IR-method and MCR-method produce reasonably similar measurementresults for wavelengths below 1 m. However, MCR-method outperforms IR-method in terms of accuracy forwavelengths exceeding 3 m. Furthermore, it was observed that IR-method, while operating at a slower speed,carries the risk of derailing and is incapable of detecting rail weld joints and low joints within the track.Originality/value – The research compare two methods’ measurement effects in a longer wavelength rangeand demonstrate the superiority of MCR-method.展开更多
社交媒体在灾害发生时能够快速提供实时且丰富的灾情信息,为应急救援提供辅助参考。然而,社交媒体信息通常以短文本形式呈现,具有口语化、语义特征稀疏和标注语料匮乏等特征,给灾情信息的识别与分析带来挑战。为此,本文提出了一种结合...社交媒体在灾害发生时能够快速提供实时且丰富的灾情信息,为应急救援提供辅助参考。然而,社交媒体信息通常以短文本形式呈现,具有口语化、语义特征稀疏和标注语料匮乏等特征,给灾情信息的识别与分析带来挑战。为此,本文提出了一种结合灾害领域知识的预训练语言模型增强方法,用于识别和分类灾情信息。首先,构建灾情知识库,包含不同灾损事件的触发词及论元;其他,通过分析短文本与灾损事件触发词的语义相似度,生成灾损知识编码;最后,将灾情领域知识与预训练词向量融合增强特征向量并输入神经网络模型实现多标签分类。以2021年7月20日前后河南暴雨灾情数据为例,将本文方法与TextCNN、Attention based CNN模型进行了对比实验,结果表明,该方法不仅有效提升了小样本数据的分类精度,还有效缓解了语义高度重合的数据类型容易错分的问题。同时,对分类结果进行灾损事件论元匹配能够充分挖掘涉灾短文本中的有效灾情信息,辅助应急救援决策。展开更多
针对手势识别由于分割效果差,导致识别率较低等问题,提出基于改进支持向量机的动态多点手势动作识别方法。选用深度阈值法分割动态多点手势图像,提取出手掌中最大的圆细化手部区域,获取7维手部HOG(Histogram of Oriented Gradients)特...针对手势识别由于分割效果差,导致识别率较低等问题,提出基于改进支持向量机的动态多点手势动作识别方法。选用深度阈值法分割动态多点手势图像,提取出手掌中最大的圆细化手部区域,获取7维手部HOG(Histogram of Oriented Gradients)特征向量,完成手势动作图像预处理。引入支持向量机,并且通过误差项改进该算法。采用改进后的支持向量机最优线性分类特征向量,利用支持向量机输入分类后的手势特征向量,实现动态多点手势动作识别。实验结果表明,所提方法受光照影响波动小,在有光照情况下,识别率达到92.5%以上,而无光照情况下,识别率仍高于90.0%,并且图像分割信息完整、识别准确性高。展开更多
基金supported in part by the National Outstanding Youth Talents Support Program(No.61822304)the Basic Science Center Program of the NSFC(No.62088101)+2 种基金the Project of Major International(Regional)Joint Research Program of NSFC(No.61720106011)the Shanghai Municipal Science and Technology Major Project(No.2021SHZDZX0100)the Shanghai Municipal Commission of Science and Technology Project(No.19511132101).
文摘The multi-point dynamic aggregation(MPDA)problem is a challenging real-world problem.In the MPDA problem,the demands of tasks keep changing with their inherent incremental rates,while a heterogeneous robot fleet is required to travel between these tasks to change the time-varying state of each task.The robots are allowed to collaborate on the same task or work separately until all tasks are completed.It is challenging to generate an effective task execution plan due to the tight coupling between robots abilities and tasks'incremental rates,and the complexity of robot collaboration.For effectiveness consideration,we use the variable length encoding to avoid redundancy in the solution space.We creatively use the adaptive large neighborhood search(ALNS)framework to solve the MPDA problem.In the proposed algorithm,high-quality initial solutions are generated through multiple problem-specific solution construction heuristics.These heuristics are also used to fix the broken solution in the novel integrated decoding-construction repair process of the ALNS framework.The results of statistical analysis by the Wilcoxon rank-sum test demonstrate that the proposed ALNS can obtain better task execution plans than some state-of-the-art algorithms in most MPDA instances.
基金Supported by the National Key R&D Program of China(2023YFB3709602,2023YFB3709603)National Natural Science Foundation of China(12372141)the Key R&D Program in Shaanxi Province(2024GH-ZDXM-27).
文摘Sandwich structures are vulnerable to multi-point impacts,and such impacts can result in a reduction in residual strength even catastrophic accident.Therefore,the multi-point impact behaviors of PMI foam sandwich structure are investigated and studied using experimental and numerical coupled methods.Three impact energy levels and five Distances Between Impact Positions(DBIP)are considered in details,and representative impact characteristics are compared to reveal the association between Compression After Impact(CAI)strength and DBIP.Results indicate that the interference between the multi-point impact events has a dominant effect on CAI strength when DBIP is small,and the variation in bending stiffness induced by the boundary effect is the dominant factor affecting CAI strength when DBIP ranges from 20 mm to 60 mm.In addition,matrix damage represents the primary damage mode in multi-point impact,and the calculated ratio of energy absorbed by the top face sheet and honeycomb core,in relation to the total absorbed energy,serves as a clear indicator of the damage severity experienced by both components.This work is enlightening for the structural design of impact-resistant composites.
文摘Purpose – Straightness measurement of rail weld joint is of essential importance to railway maintenance. Dueto the lack of efficient measurement equipment, there has been limited in-depth research on rail weld joint with a5-m wavelength range, leaving a significant knowledge gap in this field.Design/methodology/approach – In this study, the authors used the well-established inertial referencemethod (IR-method), and the state-of-the-art multi-point chord reference method (MCR-method). Two methodshave been applied in different types of rail straightness measurement trollies, respectively. These instrumentswere tested in a high-speed rail section within a certain region of China. The test results were ultimatelyvalidated through using traditional straightedge and feeler gauge methods as reference data to evaluate the railweld joint straightness within the 5-m wavelength range.Findings – The research reveals that IR-method and MCR-method produce reasonably similar measurementresults for wavelengths below 1 m. However, MCR-method outperforms IR-method in terms of accuracy forwavelengths exceeding 3 m. Furthermore, it was observed that IR-method, while operating at a slower speed,carries the risk of derailing and is incapable of detecting rail weld joints and low joints within the track.Originality/value – The research compare two methods’ measurement effects in a longer wavelength rangeand demonstrate the superiority of MCR-method.
文摘社交媒体在灾害发生时能够快速提供实时且丰富的灾情信息,为应急救援提供辅助参考。然而,社交媒体信息通常以短文本形式呈现,具有口语化、语义特征稀疏和标注语料匮乏等特征,给灾情信息的识别与分析带来挑战。为此,本文提出了一种结合灾害领域知识的预训练语言模型增强方法,用于识别和分类灾情信息。首先,构建灾情知识库,包含不同灾损事件的触发词及论元;其他,通过分析短文本与灾损事件触发词的语义相似度,生成灾损知识编码;最后,将灾情领域知识与预训练词向量融合增强特征向量并输入神经网络模型实现多标签分类。以2021年7月20日前后河南暴雨灾情数据为例,将本文方法与TextCNN、Attention based CNN模型进行了对比实验,结果表明,该方法不仅有效提升了小样本数据的分类精度,还有效缓解了语义高度重合的数据类型容易错分的问题。同时,对分类结果进行灾损事件论元匹配能够充分挖掘涉灾短文本中的有效灾情信息,辅助应急救援决策。
文摘针对手势识别由于分割效果差,导致识别率较低等问题,提出基于改进支持向量机的动态多点手势动作识别方法。选用深度阈值法分割动态多点手势图像,提取出手掌中最大的圆细化手部区域,获取7维手部HOG(Histogram of Oriented Gradients)特征向量,完成手势动作图像预处理。引入支持向量机,并且通过误差项改进该算法。采用改进后的支持向量机最优线性分类特征向量,利用支持向量机输入分类后的手势特征向量,实现动态多点手势动作识别。实验结果表明,所提方法受光照影响波动小,在有光照情况下,识别率达到92.5%以上,而无光照情况下,识别率仍高于90.0%,并且图像分割信息完整、识别准确性高。
