未知空间目标纹理缺失表面三维重建算法

针对传统三维重建方法难以对纹理缺失表面进行完整重建的问题,提出一种基于深度学习与截断符号距离函数(TSDF)融合的未知目标三维表面完整重建算法。首先设计一种基于深度学习的图像逐像素深度估计框架,通过在训练过程中引入多个复杂结构模型,提高该深度估计框架的泛化能力;其次,利用TSDF对各帧图像所估计的深度信息进行融合,实现对纹理缺失区域的空间目标完整三维重建。根据仿真校验,对于300 mm尺寸的卫星模型图像,像素深度估计平均误差约为13 mm,通过TSDF融合后尺寸精度误差小于5.10%。实验结果表明该算法可以对未知空间目标光学图像进行逐像素深度估计,并获得目标完整的三维结构与纹理信息,有效解决无纹理区域的重建结构缺失问题。     

基于光束偏转层析技术的Radon变换迭代法

本文通过计算机模拟研究,利用气体温度场的先验知识,考查了将Ｒａｄｏｎ变换迭代法用于光束偏转层析技术重建三维温度场的重建精度。作为一个应用实例,测试了电弧某截面上的温度分布。     

位相物体的三维非对称温度场的重建

采用卷积背抽影公式实现了位相物体的三维非对称温度场的重建,为光线偏／折测量三维非对称温度场提供了高精度分析和处理实验数据的方法。     

5005铝合金分流模挤压过程有限元模拟

采用DEFORM-3D有限元商业软件,利用焊合面重构技术实现了空心型材分流组合模挤压过程模拟,获取了挤压过程中的材料流动行为与流动速度场的规律,结合型材挤出时模拟的应力场和应变场的分布,研究型材挤压的晶粒分布规律并其成因。结果表明:挤压型材的横截面上的应力应变分布与实际晶粒尺寸分布有对应关系,型材挤压成形过程中应力应变集中分布是造成型材边部晶粒尺寸比中心部细小,拐角区域晶粒比直边区域细小的原因。     

绳系卫星部署阶段的动力学分析

绳系卫星系统部署阶段末时刻状态决定着其编队飞行的初始状态，为了研究不同因素对绳系卫星系统部署阶段运动的影响，本文建立了一种简单的平面哑铃模型，将系绳视作有阻尼的弹簧，两个卫星视作刚体，考虑重力梯度力矩。通过仿真，发现系绳释放的速度和副星推力是影响绳系卫星系统部署阶段稳定性的主要因素。     

基于量子行为鸽群优化的无人机紧密编队控制

随着军事、民用需求的提高和相关领域的技术推动，无人机编队协同作业已成为当今人们逐渐关注的焦点。无人机紧密编队是指无人机间侧向距离在1~2倍翼展内的编队，因其可有效改善编队中无人机的气动性能而备受瞩目。基于无人机紧密编队条件下的气动耦合效应，建立三维空间下的状态空间方程描述双机相对运动，并推导了紧密编队条件下的双机最优编队构型，在此构型下将人工势场法和编队控制相结合作为控制系统中的间接控制环，并针对基本鸽群优化算法的寻优缺陷利用量子行为规则对其进行改进，将改进后的鸽群优化算法和无人机控制量结合作为控制系统中的直接控制环，最后通过仿真对比验证了此控制系统对于紧密编队控制的有效性。     

立方体星的技术发展和应用前景

系统论述了立方体星技术发展的三个阶段;从2010-2013年已成功发射的立方体星与纳卫星中挑选了10颗典型卫星,研究分析其技术水平、应用价值和特点优势;并重点论述了立方体纳卫星的应用;最后预测了立方体纳卫星的未来应用前景。     

含噪信号经验模式分解虚假分量识别方法

针对含噪信号Hilbert-Huang变换存在虚假分量,提出改进的奇异值分解(SVD)方法进行降噪,改进包含两个部分:一是利用重构相空间代替传统矩阵如Hankel矩阵,以去掉信号冗余,再者提出奇异值能量熵分量差分法,更易于定出重构奇异值阶次;二是提出了频谱比值法对虚假分量进行辨识,更有效辨识出虚假分量.首先利用经验模式分解(EMD)得到本征模式分量(IMF),识别并剔除趋势项,重构信号,然后进行SVD,重构降噪后的信号,消除虚假分量,最后进行时频分析.联合方法应用于含噪仿真信号,信噪比(signal noise ratio,SNR)提高了5.5%,虚假分量辨识率提高至100%,用于双跨转子故障振动信号,得到正确的时频结果,表明了所提方法识别含噪信号虚假分量的有效性.     

Diversity of Comets: Formation Zones and Dynamical Paths

The past dozen years have produced a new paradigm with regard to the source regions of comets in the early solar system. It is now widely recognized that the likely source of the Jupiter-family short-period comets (those with Tisserand parameters, T > 2 and periods, P, generally < 20 years) is the Kuiper belt in the ecliptic plane beyond Neptune. In contrast, the source of the Halley-type and long-period comets (those with T < 2 and P > 20 years) appears to be the Oort cloud. However, the comets in the Oort cloud almost certainly originated elsewhere, since accretion is very inefficient at such large heliocentric distances. New dynamical studies now suggest that the source of the Oort cloud comets is the entire giant planets region from Jupiter to Neptune, rather than primarily the Uranus-Neptune region, as previously thought. Some fraction of the Oort cloud population may even be asteroidal bodies formed inside the orbit of Jupiter. These comets and asteroids underwent a complex dynamical random walk among the giant planets before they were ejected to distant orbits in the Oort cloud, with possible interesting consequences for their thermal and collisional histories. Observational evidence for diversity in cometary compositions is limited, at best. This revised version was published online in June 2006 with corrections to the Cover Date.     

Periodic relative orbits of two spacecraft subject to differential gravity and electrostatic forcing

Coulomb forces between charged close-flying satellites can be used for formation control, and constant electric potentials enable static equilibria solutions. In this work, open-loop time-varying potential functions, which produce periodic, two-craft, Coulomb formation motions are demonstrated for the first time. This is done in the rotating Hill-Frame, with linearized gravity, and craft position components assumed in the form of simple harmonic oscillators. Substitution of the oscillatory functions into the dynamics, further constrains these functions, and yields necessary potential histories, to produce the periodic flow. The assumed position functions, however, are not arbitrary, since the dynamical model restricts what oscillatory trajectories are allowed. Specifically, a Hill-Frame integral of motion is derived, and this is used to show certain candidate periodic functions to be inadmissible. The system dynamics are then linearized to expose stability properties of the solutions, and it is established that asymptotic stability is impossible for all orbit families. Finally, the degree of instability in the assumed motions, over free parameter ranges, is determined numerically via the Floquet multipliers of the associated full-cycle state-transition matrices.