【作者】 靖鑫；

【导师】 宋申民；

【作者基本信息】 哈尔滨工业大学 ， 导航制导与控制， 2010， 硕士

【摘要】 在当前空间探测活动中,交会对接技术已经不可或缺。随着空间技术的发展,卫星星座、卫星编队飞行等已在各航天大国的军事和民用技术中越来越广泛地应用,并发挥了越来越重要的作用。在这些多卫星的组合运行中,对空间目标的交会和捕获是飞行器交会对接和攻击敌对目标的重要支撑理论,是当今世界航天技术中的一项十分复杂,难度很大,并且正在迅速发展的高新技术。空间交会技术是指在空中飞行中,两个或两个以上的航天器通过轨道参数的协调,在同一时间到达同一位置的过程。空间交会过程一般包括四个阶段:远程导引段、近程寻的段、最后接近段、逼近对接阶段。其中远程导引阶段是航天器空间交会对接的重要组成阶段,此阶段往往是整个机动中能量消耗最大,时间消耗最长的阶段。而航天器携带的燃料又有限,因此对这个阶段提出优化问题是可行的。在逼近对接阶段中,近年来提出的利用双目视觉方法来测量相对位置和姿态十分有效。而在这个过程中需要用到图像匹配技术,由于图像匹配处理数据大,所以在这方面利用微粒群算法来优化其响应时间也是可行的。本文主要针对微粒群算法作了进一步的改进,并将改进算法应用于最优轨道设计和图像匹配中。主要研究内容如下:首先,介绍了基于微粒群算法的原理,给出了算法设计的基本步骤及算法流程。并对微粒群算法进行了社会行为分析。接着介绍了两种最基本的改进方法。其次,针对微粒群算法的缺点,提出了几种改进方法,并说明了它们的基本流程。再用几个标准测试函数来测试其有效性。再次,推导了Lambert定理,得到变形的Lambert定理,接着计算了目标航天器运动时间和能量消耗Δv,从而得到性能指标函数,并利用改进的微粒群算法解决最优Lambert转移。仿真实验证明了这种改进方法的有效性。最后,介绍了图像匹配的基本概念,详细讲述了harris角点检测算子和LACC角点匹配算法。考虑到实时性问题,利用微粒群算法来求得变换矩阵来减少匹配时间。更多还原

【Abstract】 In the current space exploration activities, rendezvous and docking technology is indispensable. With the development of space technology, satellite constellation and satellite formation flying are more widely used in the military and civilian technology of many country, and play an increasingly important role. Among combined operation of these multi-satellite, space rendezvous and capture is an important support theory and technology of the target spacecraft rendezvous and docking and attack enemy targets, is a very complex and difficult aerospace technology in today’s world, and is rapidly developing high-technology now.Space rendezvous technology is the process which two or more spacecraft flying in the air reach the same location at the same time through coordinating orbital parameters. Space rendezvous process generally consists of four stages: long-range guided stage, short-range homing phase, the final approach stage, approaching docking stage. Where, the long-range guided phase is an important component stage of spacecraft rendezvous and docking. This stage is the stage which often costs the biggest energy consumption and consumes longest time. It would be feasible to propose optimization problem to this phase because the fuel spacecraft carrying is limited. In the approaching docking stage, the use of binocular vision method to measure the relative position and attitude which proposed in recent years is very effective. And this process need to use the image matching technique which need process large amount of data. Therefore, it also would be feasible to make use of particle swarm algorithm to optimize its response time.This article focuses on the improvement of particles swarm optimization(PSO), and apply improved algorithm to the optimal orbit design and image matching. Main content is as following:Firstly, the principle of basic PSO algorithm is introduced, and the basic steps and procedure of PSO algorithm are proposed. The social behavior of PSO algorithm is analysised. Then two basic improvement methods are introduced.Secondly, I proposed several improved methods based on the shortcomings of PSO algorithm, and describes their basic processes. Then several standard test functions are used to test their effectiveness.Thridly, Lambert theorem is deduced, transformed Lambert theorem is acquired, and the target spacecraft moving time and energy consumptionΔv are calculated. Then the expression of objective function is getted. And use the improved PSO algorithm to solve optimal Lambert transfer. Simulation result shows that the improved method is feasible.Finally, the basic concept of image matching is introduced, and the harris corner detection operator and the LACC corner matching algorithm are detailed explained. Taking into account the real-time problem, I use PSO algorithm to obtain the transformed matrix and then reduce the match time.更多还原