一种地球紫外中心快速提取方法

针对地球紫外中心指向高精度提取问题,提出一种基于梯度统计的快速、低存储需求的紫外地平圆盘中心提取算法。首先,考虑到星载计算机与高内存消耗的滤波算法、实时导航需求的冲突,采用结合Sobel边缘算子与局部二值模式（LBP）算子的改进边缘快速提取方法,有效剔除背景噪声并准确提取地球紫外临边特征。然后对非完整临边边缘采用最小二乘拟合得到中心的精确位置。实验结果表明,该方法抗噪声,可实现亚像素级地球中心提取,并显著降低存储需求和计算时间开销。对于1046×746的图像,该算法需要的存储空间仅为1100kByte,运算时间在20ms内,满足自主导航的需求。     

Monocular Vision-based Two-stage Iterative Algorithm for Relative Position and Attitude Estimation of Docking Spacecraft

Visual sensors are used to measure the relative state of the chaser spacecraft to the target spacecraft during close range rendezvous phases. This article proposes a two-stage iterative algorithm based on an inverse projection ray approach to address the relative position and attitude estimation by using feature points and monocular vision. It consists of two stages: absolute orientation and depth recovery. In the first stage, Umeyama's algorithm is used to fit the three-dimensional (3D) model set and estimate the 3D point set while in the second stage, the depths of the observed feature points are estimated. This procedure is repeated until the result converges. Moreover, the effectiveness and convergence of the proposed algorithm are verified through theoretical analysis and mathematical simulation.     

初值不确定轨道可达区域计算

为计算航天器轨道初值不确定造成的可达区域（RD）,提出一种适用于椭圆参考轨道的计算方法。以航天器的名义轨道作为基线,定义一个随时间变化的移动参考平面,在此移动平面上建立名义轨道到可达区域上任一点的距离函数,函数的极大值点即为可达区包络的边界点,从而将问题转换为非线性方程组的求解问题,并对动力学模型部分线性化带来的误差进行了分析。随后采用变参数的自适应同伦算法对方程组进行求解。最后对航天器释放微小卫星的实例进行数值仿真,结果显示该方法可行有效。     

Lunar lander's three-dimensional translation and yaw rotation motion estimation during a descent phase using optical navigation

Optical navigation for a lunar lander consists of estimating a lander's 3-dimensional (3-D) relative dynamic motion with respect to a preselected landing site using a passive 2-dimensional (2-D) video image sequence. Lunar landing missions require a lander to perform an autonomous accurate landing with simple mechanical structure, easy operation and low cost. These requirements have motivated the need to develop an advanced navigation system. Existing navigation systems trade-off simplicity, accuracy and cost. High accuracy navigation systems typically imply complexity and high cost. In this paper, we consider a scenario where the descending phase starts from an initial altitude of 10 km with a time-of-descent of 100 s. The navigation camera is an off-the-shelf optical instrument used to take the video image sequence of the landing site during the landing phase. It is fed into the motion estimation algorithm to be processed. The continuous wavelet transform (CWT) is used to analyse each image frame of the input digital video image sequence. The output is a 2-D video image motion trajectory map, which represents the projection motion of the landing site. The 2-D video image motion is projected back to the 3-D lander's relative motion based on a geometric analysis. The outputs of this estimation algorithm are the 3-D attitude motion parameters of the lander at a time corresponding to an image being taken. The attitude determination and control system (ADCS) of the lander uses these data to perform the lander's attitude control task. In this article, we provide the motion modelling for a lunar lander during the descending phase. The projection of a 3-D planar to 2-D image plane is analysed which build the correspondence between the 3-D lander's motion and the 2-D image motion. This link provides the evidence for the geometry analysis. CWT is reviewed and CWT for video image sequence analysis is also introduced. Numerical simulation of the estimated 2-D video image sequence under the lander performing a 3-D translation and yaw rotation during the terminal descent are shown to verify the proposed concepts. The analysis of the results show that the proposed method achieves highly accurate 2-D video image motion estimation of less then 1% error with significant savings of cost, mass and volume. It leads to the accurate estimation of the lander's 3-D relative motion with respect to the landing site.     

不同工况下高速角接触球轴承-刚性转子系统动力学耦合特性分析

根据建立的高速角接触球轴承-刚性转子系统动力学数值仿真模型,以某仪表轴承支承的转子系统为算例,分析了考虑转子振动与否对系统动力学性能的影响,并详细探究了轴向载荷以及转速对转子振动、轴承内部载荷分布及旋滚比、保持架的受力和质心运动以及磨损情况的影响.结果表明:考虑转子振动时,在纯轴向载荷下,轴承各位置处的球载荷也会存在差...     

轴承支撑的舵面热模态试验及支撑刚度辨识

对在大气层内及临近空间内长时间飞行的高超声速飞行器，其舵面的模态特性比固支的翼面更加复杂，除了与舵面自身的弹性模量及内部热应力有关外，还受到根部支撑刚度的较大影响，并且支撑刚度还将受到温度的影响。以轴承机构支撑的舵面为对象，考虑温度对支撑刚度的影响，建立了非固支的全动舵面支撑边界条件。通过设计舵面受热相同、支撑部位受热不同的加热工况，辨识出了连接面两侧温升对舵面支撑刚度的线性影响规律，并验证了辨识结果的有效性。结果表明：在舵面受热相同情况下，降低支撑部位的温升，可以有效减少舵面模态频率受气动加热的影响。研究结果可供安装此类舵面的飞行器防热设计参考。     

考虑多禁止指向区域的航天器反步姿态机动控制

针对规避多个禁止指向区域的航天器姿态机动控制问题，本文在利用势函数法构造约束条件的基础上设计了一种基于反步法的航天器姿态机动控制律。首先，构造了一种通用的基于误差四元数的航天器锥形禁止指向姿态集描述，并将其转化为凸约束集形式。其次，基于该凸约束集，提出了一种物理意义明确的新型凸势函数构造方法，且该势函数仅存在唯一全局最小值，避免了传统势函数的局部极小值问题。进一步地，考虑多个禁止指向区域，设计了一种基于反步法的航天器姿态机动控制律。最后，通过典型数值仿真算例表明了本文所提出控制算法的有效性和优越性。     

Relative position and attitude estimation of spacecrafts based on dual quaternion for rendezvous and docking

The capacity to acquire the relative position and attitude information between the chaser and the target satellites in real time is one of the necessary prerequisites for the successful implementation of autonomous rendezvous and docking. This paper addresses a vision based relative position and attitude estimation algorithm for the final phase of spacecraft rendezvous and docking. By assuming that the images of feature points on the target satellite lie within the convex regions, the estimation of the relative position and attitude is converted into solving a convex optimization problem in which the dual quaternion method is employed to represent the rotational and translational transformation between the chaser body frame and the target body frame. Due to the point-to-region correspondence instead of the point-to-point correspondence is used, the proposed estimation algorithm shows good performance in robustness which is verified through computer simulations.     

基于Sarsa(λ)强化学习的空间机械臂路径规划研究

针对目标特性未知的在轨操作环境，研究了典型空间操作机械臂的路径规划策略。采用Sarsa(λ)强化学习方法实现目标跟踪及避障的自主路径规划与智能决策，该方法将机械臂系统的每节臂视为一个决策智能体，通过感知由目标偏差和障碍距离程度组成的二维状态，设计符合人工经验的拟合奖赏函数，进行各臂转动动作的强化训练，最终形成各智能体的状态-动作值函数表，即可作为机械臂在线路径规划的决策依据。将本方法应用于多自由度空间机械臂路径规划任务，仿真结果表明新算法能在有限训练次数内实现对移动目标的稳定跟踪与避障，同时各智能体通过学习所得的状态-动作值函数表，具备较强的后期在线自主调整能力，从而验证了算法较强的鲁棒性和智能性。