基于核主成分分析的铁谱磨粒特征提取方法研究

针对铁谱分析的磨粒识别过程中存在原始磨粒特征描述指标参数多、非线性突出的问题,提出基于核主成分分析的磨粒特征提取方法,介绍该方法的原理与算法。结合某柴油发动机故障检测与分析系统中铁谱磨粒自动识别的应用实例,并与传统主成份分析方法进行对比分析,结果表明该方法在进行样本非线性特征参数指标综合以及特征维数压缩方面具有可行性和有效性。     

The impact of demand uncertainty on product line design under endogenous substitution

The existing product line design literature devotes little attention to the effect of demand uncertainty. Due to demand uncertainty, the supply‐demand mismatch is inevitable which leads to different degrees of lost sales depending on the configuration of product lines. In this article, we adopt a stylized two‐segment setup with uncertain market sizes and illustrate the interplay between two effects: risk pooling that mitigates the impact of demand uncertainty and market segmentation that facilitates consumer differentiation. Compared to downward substitution, inducing bidirectional substitution through product line decisions including quality levels and prices can yield greater risk pooling effects. However, we show that the additional benefit from the risk pooling effect cannot compensate for the reduced market segmentation effect. We demonstrate that the presence of demand uncertainty can reduce the benefit of market segmentation and therefore the length of product lines in terms of the difference between products. We also propose three heuristics that separate product line and production decisions; each of these heuristics corresponds to one particular form of demand substitution. Our numerical studies indicate that the best of the three heuristics yields performance that is close to optimality. © 2015 Wiley Periodicals, Inc. Naval Research Logistics 62: 143–157, 2015     

无监督特征学习的人体活动识别

针对人的局限性可能会导致在提取特征中丢失重要信息,从而影响最终的识别效果问题,提出无监督特征学习技术的惯性传感器特征提取方法。其核心思想是使用无监督特征学习方法学习多个特征映射,再将所有特征映射拼接起来形成最终的特征计算方法。其优点是不会造成重要信息的损失,而且可以显著减少所使用的无监督特征学习模型的规模。为了验证所提出的特征提取方法在活动识别中的有效性,运用一个公开的活动识别数据集,使用三种常用无监督模型进行特征提取,并使用支持向量机进行活动识别。实验结果表明,特征提取方法取得了良好的效果,与其他方法相比具有一定的优势。     

多视图三角化中特征点噪声尺度的自适应估算

鲁棒性多视图三角化方法通常借助重投影误差经验阈值来剔除图像对应中的错误匹配,该经验阈值的选取直接影响三维重构场景点的数量和精度。在分析图像特征点定位噪声及对极传递几何原理的基础上,建立对极传递过程不确定性的传递模型,提出一种基于核密度估计的最优噪声尺度估算方法,并将该噪声尺度作为多视图三角化中错误匹配筛选的依据。实验结果表明,该方法可以获得准确的噪声尺度估计,从而有效提升多视图三角化方法的三维重构质量。     

火箭发动机故障检测的快速增量单分类支持向量机算法

为解决液体火箭发动机故障诊断正负样本不平均问题,以及实现发动机稳态工作段自适应故障检测,建立了基于快速增量单分类支持向量机的异常检测模型。采取特征工程方法,对传感器获得的多变量时间序列进行特征提取。通过增量学习方法,对单分类支持向量机模型进行改进,并应用于液体火箭发动机异常检测,使单分类支持向量机检测模型具备对不同台次、不同工况的自适应性,提高了模型的计算速度。对多台次热试车数据的分析结果表明,该模型十分有效,训练速度快,具备实用价值。     

基于直线基元角方向对几何关系直方图的目标识别方法

针对以往"角方向对几何"同类技术[6～8]提出了新的基于直线基元的几何不变特征关系,并使用该关系针对目标线集建立特征关系多维直方图,作为目标的几何不变特征,通过多维直方图匹配进行目标识别。实验证明该模式关系对于较规则的形体,其检测精度优于以往方法,可用于基于内容与结构的图像分析与目标识别系统中。     

基于KP CA的作战仿真实验数据特征提取方法

针对作战仿真分析过程中各作战要素的复杂性与非线性,研究了一种基于KPCA的作战仿真实验数据特征提取方法。该方法描述了KPCA特征提取的原理和算法,并将其应用于作战仿真实验数据的空间降维,根据累积贡献率确定新特征的数量。仿真结果表明,该方法与PCA相比具有主成份特征明显、贡献率集中等优点,能够有效综合原始数据的非线性特征,降低原始数据的维数。     

Fleet management in European integration: the case of military helicopter support

The article focuses on both economics and spatial issues related to Maintenance Repair and Overhaul (MRO) of defence platforms in Europe. Using facts and figures related to costs, trends and fleet sizes evolution, the authors discuss the interest of a spatial reorganisation of MRO production. Given the fiscal situation of the main European countries with a modern army, the relevant scale to consider the reorganisation is certainly Europe. Through the example of defence helicopters fleets of a selection of European countries, the authors address the question of cooperation of the MRO. A greater cooperation in the support of fleets would lead to a spatial reorganisation of MRO. On one side economies can be expected from this reorganisation, but, on the other hand, new problems arise. Specifically, a number of costs associated with different forms of “distances” – geographical, operational and political distances – limit the potential savings that could be expected. This work provides insights on these important issues for the construction of a European defence, both in its political, military and industrial dimensions.     

一种基于模型的舰船目标特征提取方法

针对中分辨雷达,提出了一种基于模型的舰船目标特征提取方法。通过对特征向量的组合和加权提高了舰船目标的特征提取质量,在人工神经网络识别器中取得了较好的识别效果。     

Inventory pooling games for expensive,low‐demand spare parts

We consider several independent decision makers who stock expensive, low‐demand spare parts for their high‐tech machines. They can collaborate by full pooling of their inventories via free transshipments. We examine the stability of such pooling arrangements, and we address the issue of fairly distributing the collective holding and downtime costs over the participants, by applying concepts from cooperative game theory. We consider two settings: one where each party maintains a predetermined stocking level and one where base stock levels are optimized. For the setting with fixed stocking levels, we unravel the possibly conflicting effects of implementing a full pooling arrangement and study these effects separately to establish intuitive conditions for existence of a stable cost allocation. For the setting with optimized stocking levels, we provide a simple proportional rule that accomplishes a population monotonic allocation scheme if downtime costs are symmetric among participants. Although our whole analysis is motivated by spare parts applications, all results are also applicable to other pooled resource systems of which the steady‐state behavior is equivalent to that of an Erlang loss system. © 2012 Wiley Periodicals, Inc. Naval Research Logistics, 2012