湍动能模拟内部声激励的计算研究

为了研究襟翼表面的声激励对多段翼型升力特性的影响机理,利用湍动能假设进行数值计算研究。以湍动能的量值、雷诺数以及襟翼与主翼的搭接量为参数,研究上述参数的变化对升力特性的影响,得到的结果与实验值符合良好。在多段翼型缝道处加入湍动能的影响规律是：（1）在升力特性线性段减少了升力系数,在失速点附近可以推迟分离,提高升力系数;（2）当雷诺数增加时,湍动能对升力系数的影响量减小;（3）在搭接量为零时,湍动能的影响最大。上述规律与在缝道处加入声学激励的影响规律一致,表明采用注入湍动能来类比声学激励有一定的应用价值。     

Wire electrochemical machining with axial electrolyte flushing for titanium alloy

Titanium and its alloys have found very wide application in aerospace due to their excellent characteristics although their processing is still a challenge. Electrochemical machining is an important issue in the fabrication of titanium and titanium alloys. Wire electrochemical machining (WECM) is mainly used for workpiece cutting under the condition of different thickness plates. It has a great advantage over wire electro-discharge machining, which is the absence of heat-affected zone around the cutting area. Moreover, the wire electrode in WECM could be used repetitively because it is not worn out. Thus, much attention has been paid to WECM. The effective way of removing electrolysis products is of importance to WECM. In this paper, the axial electrolyte flushing is presented to WECM for removing electrolysis products and renewing electrolyte. The Taguchi experiment is conducted to optimize the machining parameters, such as wire feedrate, machining voltage, electrolyte concentration, etc. Experimental results show that WECM with axial electrolyte flushing is a promising issue in the fabrication of titanium alloy (TC1). The feasibility of multi-wire electrochemical machining is also demonstrated to improve the machining productivity of WECM.     

C/ SiC 材料在国外空间光学系统上的应用

C/SiC复合材料具有高的比强度、低的热变形敏感度以及在高低温环境下的适应性,这使其成为目前最具前途的空间光学系统应用材料.本文总结了国外一些国家C/SiC的制备方法及其在空间技术上的具体应用.     

基于高效垂直互连的X频段三维交叉合路网络

针对射频微波系统小型化、一体化、低成本设计需求,利用HFSS软件3D建模仿真研究微波毫米波多层板高密度垂直互连技术,对比不同结构参数的频率特性,在结构上通过加载层间焊盘改善特定频段内传输性能,在DC^20GHz内回波损耗小于–20dB。基于该高效垂直互连技术,实现32路信号输入4波束输出交叉网络3D垂直合成,体积仅为125mm×30mm×1.8mm,经测试带内插损≤10.95dB,驻波比≤1.4,较好地实现了X频段合路输出功能。     

红外可见光双目系统像素级快速融合方法

红外 可见光双目系统既能感知目标的温度，又能获取目标的颜色、纹理信息，在日常生活和工业产生中获得较多的应用.红外相机由于和可见光相机成像机制的不同，在标定过程中往往需要使用不同的靶标.本文提出了一种红外 可见光双目系统像素级的快速标定融合方法，通过设计了一种圆孔形靶标，实现了一步法标定，在可见光相机像素坐标系下像素质心标定精度小于1个像素，可用于大批量的低成本的民用红外标定产品系统.     

Preliminary analysis of the quality and positioning performance of BDS-3 global interoperable signal B1C&B2a

BeiDou-3 Navigation Satellite System (BDS-3) satellites are equipped with the new generation GNSS signals B1C and B2a, which support the interoperability with GPS and Galileo systems. In this study, the pseudo-range multipath error and carrier phase observation noise of the BDS-3 B1C and B2a signals were evaluated based on zero baseline measurements from the day of year (DOY) 113 to 116, 2020. Further, the precision and performance of the single point positioning (SPP) and precise point positioning (PPP) are assessed at 9 stations. This assessment manifests that the standard deviations (STDs) of the pseudo-range multipath error are about 0.09 ~ 0.22 m, while STDs of the carrier phase observation noise are about 0.075 mm. For the single-frequency SPP, its positioning precision is about 2.03 ~ 4.85 m and 3.29 ~ 10.73 m at the 99.99% confidence level in horizontal and vertical directions, respectively, while the dual-frequency SPP precision is about 1.92 ~ 8.02 m and 4.81 ~ 12.77 m in horizontal and vertical directions, respectively. For the daily static PPP, the convergence time is about 20 ~ 30 min, while the daily positioning precision can reach 1.38 ~ 4.42 cm and -1.31 ~ 4.34 cm in horizontal and vertical directions, respectively. In general, the quality and the SPP and PPP performance of the BDS-3 B1C&B2a signals are comparable to the GPS and Galileo.     

Inertial aided BDS triple-frequency integer ambiguity rounding method

The integer ambiguity resolution (AR) of carrier phase is significant for Global Navigation Satellite System (GNSS) precise positioning. However, in kinematic case, single-epoch AR methods based on alone GNSS are usually not reliable due to the instable pseudorange accuracy. Moreover, the computation of classical AR method Least Squares Ambiguity Decorrelation Adjustment (LAMBDA) is large. Thus, the inertial measurement unit (IMU) is introduced, a new inertial-aided AR method that directly rounds the float ambiguity of BeiDou triple-frequency combined observations, which is characterized by long wavelength, low carrier-phase noise and ionospheric delay, is proposed. The mathematical model of the new method is derived first. Then the impacts of the carrier-phase noise, ionospheric delay and inertial navigation system (INS) position error on the AR success ratio of combined observation are analyzed through probabilistic approach. Based on above investigation, the combinations (0, ?1, 1), (1, 4, ?5) and (4, ?2, ?3) are selected to resolve the original ambiguity. A vehicular integrated navigation test is performed to demonstrate the proposed method. The results show that the average AR success ratios of the three selected combinations, whose float ambiguity errors are 0.041, 0.146, 0.279 cycle respectively, are above 97.25% without regard to low-elevation C05. With respect to positioning accuracy based on our AR method when compared with IE software, the east, north, up error RMS of position are 0.042, 0.024, 0.069 m, respectively. In terms of the AR recover after the BeiDou signals outage, as long as 62 s BeiDou signal complete outage, all the ambiguities of all satellites could be re-fixed immediately. Besides, during the 90 s signals partial outage, the AR is not influenced by the position error, since the float ambiguity errors are all below half-cycle. The research of this contribution demonstrates the effectiveness of the proposed new method, which indicates it is applicable to kinematic positioning, even in BDS degraded and denied environments.     

局部Bagging方法及其在人脸识别中的应用

Bagging方法无法在稳定的分类器(如最近邻分类器)上构建多样的分量分类器,因此它不适合于稳定的分类器.同时,小样本特性Bagging也很难应用于人脸识别等任务中.本文提出了一种局部Bagging(L-Bagging)方法以同时解决上述两个问题.L-Bagging和Bagging的主要区别是L-Bagging在每个事先划分好的局部区域内进行自助集的采样而不是如Bagging那样在原始的样本集上采样.由于局部区域的维数通常远远小于训练样本数,并且分量分类器又是构建在不同的局部区域上的, 因此 L-Bagging方法不仅有效地解决了小样本问题,而且产生了更多样的分量分类器.在4个标准的人脸数据库(AR,Yale,ORL和Yale B)上的实验结果表明所提出的L-Bagging方法是有效的,并且对光照、遮挡及轻微的姿态变化是鲁棒的.     

基于分布式模型预测控制的多UAV协同区域搜索

针对多无人机(UAV)协同区域搜索问题展开研究。提出了一种基于分布式模型预测控制(DMPC)的多UAV分布式优化搜索方法。首先基于传统的搜索图模型,建立了多UAV协同搜索的问题描述和状态空间模型,然后在DMPC框架下,将集中式多UAV在线优化决策问题转化为各架UAV的小规模分布式优化问题,采用基于纳什最优和粒子群优化(PSO)相结合的算法实现对每个子系统优化问题的迭代求解。仿真结果表明:DMPC方法能够有效地降低多UAV协同搜索决策问题的求解规模,是一种可行的方法。