退化速率跟踪粒子滤波在剩余使用寿命预测中的应用

毋庸置疑,剩余使用寿命预测对于设备的健康管理越来越重要。近年来粒子滤波方法被越来越多地应用到设备寿命预测技术当中,这是因为粒子滤波方法能更好地解决非线性非高斯系统滤波问题,而且能够获得不确定度信息。但该方法的预测性能却过度依赖于预测模型,并且对于模型参数的初始分布也比较敏感,这在一定程度上限制了粒子滤波预测方法的进一步发展。针对基本粒子滤波预测方法的不足,提出了一种基于退化速率跟踪粒子滤波的通用预测框架,以历史观测数据的退化速率统计规律作为指导来跟踪目标数据的退化速率,实现对粒子滤波预测方法的简化,并将该方法用于轴承和锂离子电池的剩余使用寿命预测,验证了方法的有效性。     

Note: Two rendezvous search problems on the line

A new upper bound is obtained for the two‐person symmetric rendezvous value on the real line when the distribution function of their initial distance apart is bounded. A second result shows that if three players are placed randomly on adjacent integers on the real line facing in random directions and able to move at a speed of at most 1, then they can ensure a three‐way meeting time of at most 7/2; the fact that 7/2 is a best possible result follows from work already in the literature. © 1999 John Wiley & Sons, Inc. Naval Research Logistics 46: 335–340, 1999     

Single machine sequencing with linear models of release dates

The paper deals with a problem of scheduling a set of jobs on a single machine. Before a job is released for processing, it must undergo some preprocessing treatment that consumes resources. It is assumed that the release date of a job is a linear decreasing continuous function of the amount of a locally and globally constrained, continuously divisible resource (e.g., energy, catalyzer, financial outlay, gas). The problem is to find a sequence of jobs and a resource allocation that will minimize the maximum job completion time. Such a problem appears, for example, in the ingot preheating and hot-rolling process in steel mills. It is shown that the problem is strongly NP-hard. Some polynomially solvable cases of the problem and approximate algorithms with both experimental and worst-case analysis are presented. © 1998 John Wiley & Sons, Inc. Naval Research Logistics 45: 99–113, 1998     

Attrition models of the Ardennes campaign

This paper revisits the modeling by Bracken [3] of the Ardennes campaign of World War II using the Lanchester equations. It revises and extends that analysis in a number of ways: (1) It more accurately fits the model parameters using linear regression; (2) it considers the data from the entire campaign; and (3) it adds in air sortie data. In contrast to previous results, it concludes by showing that neither the Lanchester linear or Lanchester square laws fit the data. A new form of the Lanchester equations emerges with a physical interpretation. © 1998 John Wiley & Sons, Inc. Naval Research Logistics 45: 1–22, 1998     

不确定失效阈值影响下考虑设备剩余寿命预测信息的最优替换策略

为了进一步提升设备维修决策的科学性,通过建立综合设备剩余寿命预测数据与不确定失效阈值的最优维修决策模型,实现了不可维修设备的最优替换策略.构建基于非线性Wiener过程的设备性能退化模型,并采用极大似然法估计退化模型参数;提出一种基于期望最大(Expectation Maximization,EM)算法的不确定失效阈值...     

基于线特征的多时相遥感图像变化检测

多时相遥感图像中利用线特征可以感知道路等线形地物的变化。论文提出一种新的基于边缘特征描述的线特征变化检测算法,首先根据边缘的梯度信息得到变化边缘,并对变化边缘进行编组,形成变化边缘的直线支撑区域,然后通过最小二乘拟合方法从直线支撑区域中提取出变化直线。提出的线特征变化检测算法把边缘的相位和幅度信息作为变化检测的判定依据,从而避免了线的匹配与比较工作,降低了变化检测的算法复杂度,具有很强的实用性。将提出的方法用于多时相遥感图像变化检测,实验结果显示该方法的有效性。     

基于位置预测的靶场图像实时判读方法

在靶场经纬仪对目标实时跟踪测量时,会发生相机随机抖动的情况,引起目标在图像中大幅度运动。应对大幅度运动时,基于搜索窗口的跟踪方法容易丢失目标,而基于全图搜索的跟踪方法时效性差。针对以上问题,提出一种结合核相关滤波算法(Kernelized Correlation Filter, KCF)和目标位置预测的改进的跟踪学习检测算法(Tracking-Learning-Detection, TLD)跟踪框架。利用正交多项式最优线性滤波器及相机角度信息预测目标下一帧位置,在此区域利用KCF进行快速跟踪,可以提高跟踪的成功率和时效性,跟踪失败时再进行检测。仿真实验表明,最优线性滤波器能较准确预测目标位置,给KCF提供较准确的搜索位置,算法每帧耗时仅为1.1 ms,且定位精度优于TLD和KCF,能有效应对相机抖动的问题。靶场实际试验证明该方法可提高靶场自动判读水平,减少人工干预。     

深度神经网络指导下的无线覆盖预测算法

为保证新一代移动无线网络能够根据实时覆盖情况动态地调节小区天线参数,需要实现高效且准确的无线覆盖预测。传统的求解方法通过精确的场强预测判断天线参数的优劣,虽然精度很高但需要大量的计算资源,无法满足5G和后5G移动网络通过实时覆盖预测进行射频参数动态调整的实际需求。现采用基于深度神经网络的算法对给定天线参数的覆盖效果进行预测,以取代对目标区域的精确场强预测。数值结果表明:该方法能够在保持计算准确性的同时显著减少计算量,为5G动态网络规划提供基础性参考数据。     

磁偶极子信号检测和参数估计

在分析了经典磁偶极子模型的基础上 ,指出该模型是两个函数的组合 ,目标参数包含在这两个函数的参数里 .提出了二次函数组合法 ,能够在绝大多数情况下将这两个函数分离 ,逐步求出目标参数 .最后对实例进行了仿真计算 ,给出了较满意的结果     

一种慢波自由电子激光的动力学理论分析

本文用动力学耦合模理论,研究在矩形波导中单侧填充介质而形成的慢波线极化自由电子激光的特性。结果表明,在这一系统中,可用低能电子束产生高频率高增益的自由电子激光,并具有极宽的调谐域。