应用引导积分梯度的对抗样本生成

给图片添加特定扰动可以生成对抗样本, 误导深度神经网络输出错误结果, 更加强力的攻击方法可以促进网络模型安全性和鲁棒性的研究. 攻击方法分为白盒攻击和黑盒攻击, 对抗样本的迁移性可以借已知模型生成结果来攻击其他黑盒模型. 基于直线积分梯度的攻击TAIG-S可以生成具有较强迁移性的样本, 但是在直线路径中会受噪声影响, 叠加与预测结果无关的像素梯度, 影响了攻击成功率. 所提出的Guided-TAIG方法引入引导积分梯度, 在每一段积分路径计算上采用自适应调整的方式, 纠正绝对值较低的部分像素值, 并且在一定区间内寻找下一步的起点, 规避了无意义的梯度噪声累积. 基于ImageNet数据集上的实验表明, Guided-TAIG在CNN和Transformer架构模型上的白盒攻击性能均优于FGSM、C&W、TAIG-S等方法, 并且制作的扰动更小, 黑盒模式下迁移攻击性能更强, 表明了所提方法的有效性.     

A path planning model based on nested regular hexagons using weighted centroid localization

In many wireless sensor network (WSN) applications, the location of a sensor node is crucial for determining where the event or situation of interest occurred. Therefore, localization is one of the critical challenges in WSNs. Mobile anchor node assisted localization (MANAL) is one of the promising solutions for the localization of statically deployed sensors. The main problem in MANAL localization is that the path planning of the mobile anchor (MA) node should be done so that the localization error in the network will be minimal and that all unknown nodes in the network are covered. This paper proposes a new path planning approach called nested hexagons curves (NHexCurves) for MANAL. NHexCurves guarantees that it will receive messages from at least three non-collinear anchors to locate all unknown nodes in the network. The proposed model has compared six different path planning schemes in the literature using weighted centroid localization (WCL). In these comparisons, first of all, localization errors of the models are compared using some statistical concepts. Second, the variation of the localization error according to parameters such as resolution (R) and the standard deviation of noise (σ) is observed. Then, with similar approaches, the standard deviation of errors, localization ratio, scalability performances, and finally, path lengths of the models are examined. The simulation results show that the NHexCurves static path planning model proposed in this study stands out compared to other models with high localization error and localization ratio performance, especially at low resolutions, due to its path design. At the same time, the lowest error values according to σ are obtained with the proposed model among all models considered.     

可控震源重锤-活塞杆动态行为规律与扰动控制研究

针对工程应用中可控震源振动器扰动量大、信噪比低等问题,结合重锤运动轨迹检测试验,通过理论计算与数值模拟,开展了重锤-活塞杆动态行为规律研究;在此基础上,开展振动器扰动控制研究,提出竖直进油方式振动器,与横向进油时的结果开展对比仿真分析,并进行竖直进油振动器扰动试验验证。结果表明:横向进油方式下,重锤受绕z轴的转动力矩作用且活塞杆存在变形,导致重锤上某点的运动轨迹为空间曲线,增大了振动器的扰动量,因此,横向进油方式是导致振动器扰动大的直接原因;竖直进油方式下,重锤不再受转动力矩影响且活塞杆变形得到极大改善,有效降低了振动器扰动量。研究结果对改善输出信号品质、提高输出信号分辨率具有重要的工程意义,建议在振动器的设计中采用竖直进油方式。     

Irregular Time‐Varying Stress Degradation Path Modeling: a Case Study on Lithium‐ion Cell Degradation

The aim of this paper is to investigate the issue of degradation modeling and reliability assessment for products under irregular time‐varying stresses. Conventional degradation models have been extensively used in the relevant literature to characterize degradation processes under deterministic stresses. However, the time‐varying stress, which may affect degradation processes, widely exists in field conditions. This paper extends the general degradation‐path model by considering the effects of time‐varying stresses. The new degradation‐path model captures influences of varying stresses on performance characteristics. A nonlinear least square method is used to estimate the unknown parameters of the proposed model. A bootstrap algorithm is adopted for computing the confidence intervals of the mean time to failure and percentiles of the failure‐time distribution. Finally, a case study of lithium‐ion cells is presented to validate the proposed method. Copyright © 2015 John Wiley & Sons, Ltd.     

曲面复合材料缝合机器人自动路径规划技术

为提高复合材料预制件缝合机器人的灵活性与加工柔性,针对缝合机器人自动路径规划问题,研究了基于三维光栅扫描技术的曲面复合材料缝合机器人视觉接缝提取技术.针对三维扫描摄像机视场固定的特点,对摄像机进行标定实验,改进传统的手眼标定方法,提出一种用于求解手眼关系矩阵的四点标定新方法,建立了机器人坐标系与摄像机坐标系的位姿映射关系.经过手眼关系标定,三维光栅扫描仪采样得到的缝合预制件点云数据可转化为机器人坐标系下的点云数据坐标.进一步对采集的预制件点云数据进行接缝中心线提取,提出了一种缝合轨迹及姿态规划算法,采用三次多项式对接缝中心线进行空间曲线拟合,采用最小二乘法对缝合微切平面进行平面拟合,完成机器人操作空间前进矢量及接近矢量的计算.最后,将规划结果应用于缝合机器人的离线编程仿真与实验中.实验结果表明,该系统精度高,线迹成型良好,可以满足机器人缝合系统的要求.     

Effects of microstructure on fatigue crack propagation behavior in a bi-modal TC11 titanium alloy fabricated via laser additive manufacturing

In this study, the crack propagation behaviors in the equiaxed and equiaxed-columnar grain regions of a heat-treated laser additive manufacturing (LAM) TC11 alloy with a special bi-modal microstructure are investigated. The results indicate that the alloy presents a special bi-modal microstructure that comprises a fork-like primary α (α_p) phase surrounded by a secondary α colony (α_s) in the β phase matrix after the heat treatment is completed. The samples demonstrate a fast crack growth rate with larger da/dN values through the equiaxed grain sample versus across the equiaxed-columnar grain sample at low ΔK values (<13.8). The differences that are observed between the crack propagation behaviors (in the crack initiation stage) of the samples can be mostly attributed to the different size and morphology of the α_p lamellae and α_s colony within the grains in the equiaxed and columnar grain regions rather than the grain boundaries. The cracks prefer to grow along the α/β boundary with a smooth propagation route and a fast propagation rate in the equiaxed grain region, where the α_p and α clusters have a large size. However, in the columnar grain region, small and randomly distributed α_p lamellae generate a zigzag-shaped propagation path with a reduction in the da/dN value. Additionally, the change in the size of the α_p lamellae in the equiaxed grains (heat affected bands, HAB) is also observed to influence the propagation behavior of the crack during the crack initiation stage.     

大型差异数据库中入侵路径恢复模型分析

在大型差异数据库中,假设入侵特征存在较大的伪装,将难以形成入侵判断的依据,无法建立入侵路径恢复模型;提出基于敏感性数据挖掘方法的大型差异数据库中入侵路径恢复方法模型;利用主成分分析方法搜索大型差异数据路中的入侵路径,为模型的建立提供准确的数据基础,利用敏感性数据挖掘方法能够检测到入侵路径的特征,从而能够建立准确的入侵路径模型;实验结果表明,利用改进方法进行大型差异数据库中入侵路径恢复,能够提高恢复的准确性,从而保证大型差异数据库的安全。     

Frenet坐标框架下的自适应避障算法

目的 解决包装车间无人运输车辆在沿全局路径行驶过程中,难以同时保证跟随轨迹平滑及规避障碍物的问题。方法 构建Frenet坐标框架,描述车辆位姿与全局路径的相对关系,利用五次多项式生成多条待选局部路径;采用障碍物势场法模糊处理包装车间障碍的外形轮廓,并设置膨胀区间;结合待选路径采样点经过障碍物膨胀区间得到的碰撞值,以及采样点与全局规划路径的偏差值,综合评估路径的避障能力,筛选出最优的避障路径。结果 该算法能根据障碍物的大小,动态解算出与全局路径偏移量小且运动连贯的局部避障路径。结论 所提算法使得车辆完成运输任务时的效果更好,且效率更高。     

地下矿山巷道运输事故的致因分析

针对地下矿山巷道运输事故的案例,分析了导致事故的基本事件;通过对22个基本事件构成事故破坏模式的调查分析,得出了巷道运输事故的事故树;采用最小割集、最小径集以及结构重要度的意义对发生事故的基本事件进行定性分析,找出了导致运输事故基本事件的影响程度。该文有助于在防范事故发生时抓住主次因素,对矿山企业运输安全管理具有参考意义。     

Global optimization of tool path for five-axis flank milling with a conical cutter

In this paper, optimum positioning of the conical cutter for five-axis flank milling of slender surfaces is addressed from the perspective of approximating the tool envelope surface to the data points on the design surface following the minimum zone criterion recommended by ANSI and ISO standards for tolerance evaluation. Based on the observation that a conical surface can be treated as a canal surface, i.e. envelope surface of one-parameter family of spheres, the swept envelope of a conical cutter is represented as a sphere-swept surface. Then, an approach is presented to efficiently compute the signed distance between a point in space and the swept surface without constructing the swept surface itself. The first order differential increment of the signed point-to-surface distance with respect to the differential deformation of the tool axis trajectory surface is derived. By using the distance function, tool path optimizations for semi-finish and finish millings are formulated as two constrained optimization problems in a unified framework, and a sequential approximation algorithm along with a hierarchical algorithmic structure is developed for the optimization. Numerical examples are given to confirm the validity and efficiency of the proposed approach. Comparing with the existing approaches, the present one improves the machining accuracy greatly. The rationale developed applies to general rotary cutters.