In this paper, we propose a new formula of the real-time minimum safety headway based on the relative velocity of consecutive trains and present a dynamic model of high-speed passenger train movements in the rail line...In this paper, we propose a new formula of the real-time minimum safety headway based on the relative velocity of consecutive trains and present a dynamic model of high-speed passenger train movements in the rail line based on the proposed formula of the minimum safety headway. Moreover, we provide the control strategies of the high-speed passenger train operations based on the proposed formula of the real-time minimum safety headway and the dynamic model of highspeed passenger train movements. The simulation results demonstrate that the proposed control strategies of the passenger train operations can greatly reduce the delay propagation in the high-speed rail line when a random delay occurs.展开更多
Adaptive Delaunay triangulation is combined with the cell-centered upwinding algorithm to analyze inviscid high-speed compressible flow problems. The multidimensional dissipation scheme was developed and included in t...Adaptive Delaunay triangulation is combined with the cell-centered upwinding algorithm to analyze inviscid high-speed compressible flow problems. The multidimensional dissipation scheme was developed and included in the upwinding algorithm for unstructured triangular meshes to improve the computed shock wave resolution. The solution accuracy is further improved by coupling an error estimation procedure to a remeshing algorithm that generates small elements in regions with large change of solution gradients, and at the same time, larger elements in other regions. The proposed scheme is further extended to achieve higher-order spatial and temporal solution accuracy. Efficiency of the combined procedure is evaluated by analyzing supersonic shocks and shock propagation behaviors for both the steady and unsteady high-speed compressible flows.展开更多
战斗部破片群运动参数对弹药毁伤威力评估具有重要的意义。针对破片尺寸较小、背景信息复杂以及破片数据样本少导致的破片检测精度较低的问题,本文提出一种YOLOv5-FD的战斗部破片群目标检测方法。首先,在网络输出端增加微小目标检测层,...战斗部破片群运动参数对弹药毁伤威力评估具有重要的意义。针对破片尺寸较小、背景信息复杂以及破片数据样本少导致的破片检测精度较低的问题,本文提出一种YOLOv5-FD的战斗部破片群目标检测方法。首先,在网络输出端增加微小目标检测层,将原始的三尺度改为四尺度,并在特征融合网络中引入内容感知特征重组(Content Aware ReAssembly of Features,CARAFE)上采样模块替换原有的最近邻插值上采样,减少小目标特征信息损失,提高弱小破片的提取能力。其次,在特征提取网络引入坐标注意力模块(Coordinate Attention,CA),加强对破片特征的提取,弱化背景信息,抑制复杂背景的干扰。最后,在模型训练过程中引入模型不可知元学习方法(Model Agnostic Meta Learning,MAML),达到仅用小样本破片数据集实现较高的检测性能。实验结果表明,YOLOv5-FD破片检测算法在自制破片数据集中,精确率达到了90.5%,召回率达到了85.4%,平均精度mAP_0.5达到了88.2%,与原始YOLOv5s算法相比分别提高了7.1%,7.9%和7.5%,有效提高了破片目标检测准确性。展开更多
基金supported by the National Basic Research Program of China (Grant No. 2012CB725400)the National Natural Science Foundation of China (Grant No. 71131001-1)the Research Foundation of State Key Laboratory of Rail Traffic Control and Safety,Beijing Jiaotong University,China (Grant Nos. RCS2012ZZ001 and RCS2012ZT001)
文摘In this paper, we propose a new formula of the real-time minimum safety headway based on the relative velocity of consecutive trains and present a dynamic model of high-speed passenger train movements in the rail line based on the proposed formula of the minimum safety headway. Moreover, we provide the control strategies of the high-speed passenger train operations based on the proposed formula of the real-time minimum safety headway and the dynamic model of highspeed passenger train movements. The simulation results demonstrate that the proposed control strategies of the passenger train operations can greatly reduce the delay propagation in the high-speed rail line when a random delay occurs.
文摘Adaptive Delaunay triangulation is combined with the cell-centered upwinding algorithm to analyze inviscid high-speed compressible flow problems. The multidimensional dissipation scheme was developed and included in the upwinding algorithm for unstructured triangular meshes to improve the computed shock wave resolution. The solution accuracy is further improved by coupling an error estimation procedure to a remeshing algorithm that generates small elements in regions with large change of solution gradients, and at the same time, larger elements in other regions. The proposed scheme is further extended to achieve higher-order spatial and temporal solution accuracy. Efficiency of the combined procedure is evaluated by analyzing supersonic shocks and shock propagation behaviors for both the steady and unsteady high-speed compressible flows.
文摘战斗部破片群运动参数对弹药毁伤威力评估具有重要的意义。针对破片尺寸较小、背景信息复杂以及破片数据样本少导致的破片检测精度较低的问题,本文提出一种YOLOv5-FD的战斗部破片群目标检测方法。首先,在网络输出端增加微小目标检测层,将原始的三尺度改为四尺度,并在特征融合网络中引入内容感知特征重组(Content Aware ReAssembly of Features,CARAFE)上采样模块替换原有的最近邻插值上采样,减少小目标特征信息损失,提高弱小破片的提取能力。其次,在特征提取网络引入坐标注意力模块(Coordinate Attention,CA),加强对破片特征的提取,弱化背景信息,抑制复杂背景的干扰。最后,在模型训练过程中引入模型不可知元学习方法(Model Agnostic Meta Learning,MAML),达到仅用小样本破片数据集实现较高的检测性能。实验结果表明,YOLOv5-FD破片检测算法在自制破片数据集中,精确率达到了90.5%,召回率达到了85.4%,平均精度mAP_0.5达到了88.2%,与原始YOLOv5s算法相比分别提高了7.1%,7.9%和7.5%,有效提高了破片目标检测准确性。
