【作者】 姚克明；

【导师】 陆宇平；

【作者基本信息】 南京航空航天大学 ， 导航、制导与控制， 2011， 博士

【摘要】 随着人类社会经济与科技的发展,深空探测越来越受到各国的重视。各国空间探测技术的水平也直接影响其国际地位。火星是离地球最近的最有可能存在生命的行星。因此,火星探测成为近年来国际深空探测的热点问题。利用无人机探测火星具有其它探测方法不具备的优点,这使得火星无人机探测成为未来深空探测的发展方向之一。国外对火星无人机做了一些研究,但未见有成功试飞的报道。目前,国内尚未见火星无人机相关的研究性文献或报道。本文针对火星无人机及其探测飞行中的几个关键问题做了一些研究和探讨。本文首先描述了火星探测无人机的总体任务,研究了火星的飞行环境特点,比较了其与地球相关环境的差异,并分析了该差异对火星无人机相关研究的影响。在此基础上,本文建立了火星无人机的纵向数学模型,并对该模型进行了相关的气动特性分析。这些内容将为今后对火星无人机导航和控制方法的研究提供平台。其次,本文研究了探测无人机从出舱到平飞的整个过程特点,并根据上述模型,提出了火星无人机的切换控制方法,拉起和平飞控制方法及总能量控制方法。由于火星无人机探测飞行的任务特点决定了它的一些控制动作需要导航信号的触发,同时火星无人机本身必须具备自主导航能力,因此,本文引入了对采用计算机视觉技术进行探测无人机自主导航的相关研究。本文根据探测无人机的机载计算机的局限性及其探测飞行的特点,针对性地研究了视觉导航中的景象匹配问题和机载数据存储问题,提出了火星无人机的快速图像特征提取与景象匹配方案及其具体算法,并给出了相应的机载数据存储方法。具体而言,本文针对探测无人机的平飞段和末段分别采用不同的图像特征提取和图像匹配算法,从而实现其全程景象匹配的快速性和准确性。最后,本文对以上的研究工作进行了总结,指出了其中的不足,并给出了下一步研究工作的建议。更多还原

【Abstract】 With the development of human socio-economy and technology, more and more people from various countries pay more attention to deep-space exploration. The level of national space exploration technology has a direct impact on its international status. Mars is the closest planet where life is possible. So , Mars exploration is a hot issue in international deep space exploration in recent years. Mars UAV has the advantages that the other exploration methods do not have. It makes that Mars UAV exploration is becoming one of the development directions of deep space exploration .Some researches are done for Mars UAV in other countries, But , there is no news about the successful test flight of Mars UAV. The other documents about Mars UAV research have not been seen or been reported in China. In this paper, some research and discussion are done for the several key issues about Mars UAV.First, the overall mission of Mars UAV exploration and the feature of Mars environment surroundings are both described in this paper, The environmental differences between the Earth and Mars are compared ,and the impact on the researches for Mars UAV is analyzed. Based on the above, the longitudinal mathematical model of Mars UAV is established. And the related aerodynamic features of this model are also analyzed. These work will provide the platform for the follow-up research for Mars UAV navigation and control methods.Secondly, the feature of the whole process from extravehicular to level flight is researched. Based on the above model, two control methods for Mars UAV are proposed, one is a switching control method, the other is the method for pull-up control and level flight control . And then, a total energy control method is also proposed. Mars UAV control actions need navigation signals as the trigger signal of the effective implementation according to its task features. But, Mars UAV itself must have autonomous navigation capability. Therefore, computer vision technology is researched for autonomous navigation of Mars UAV in this paper. Scene matching and on-board data storage problem in visual navigation are researched, according to the features of Mars UAV exploration flight and the limitations of its on-board computer. Fast image feature extraction and matching algorithms for Mars UAV are proposed. Its On-board data storage mode is also proposed. Different methods are designed for different flight phases to reduce general performance requirements of its onboard computer, while achieving fast image matching of the whole Mars UAV exploration flight.Finally, the research work on the the above is summarized, and its shortcomings are also pointed out.Then,the further research proposals are given.更多还原