【作者】 张超；

【导师】 张泽旭；

【作者基本信息】 哈尔滨工业大学 ， 飞行器设计， 2009， 硕士

【摘要】 随着人类对火星探测的不断深入,将会促使越来越多的具有更高科考价值的复杂地形进入探测范围,这就对着陆器的安全软着陆提出了更高的技术要求,具有高度的障碍检测能力和障碍规避能力的软着陆自主导航技术已经成为未来火星探测的重要前瞻性技术之一。本文参考“凤凰号”火星探测器的着陆过程,对火星软着陆障碍检测与规避技术进行研究,并进行相关的仿真演示平台的开发。首先,基于CCD着陆相机采集的着陆区图像序列,分别对陨石坑和岩石障碍进行检测与识别。对陨石坑障碍,给出一种基于GVF-Snake动态模型的可靠闭合轮廓检测方法;对岩石障碍,给出一种基于多分辨率分析的快速岩石检测方法。对检测出的障碍,通过C-均值聚类方法计算出障碍目标大小及分布,并利用螺旋式搜索得到安全着陆点。其次,根据检测出的安全着陆点,对障碍规避方法进行研究,给出了基于多项式求解两点边值问题的障碍规避方法和基于李雅普诺夫稳定理论的障碍规避方法。基于李雅普诺夫稳定理论的障碍规避方法考虑到地形威胁,将李雅普诺夫函数设计成状态函数与地形威胁势函数的折衷函数,进而求得探测器障碍规避控制律。最后,在虚拟仿真引擎Delta3D开发环境下,对软着陆障碍检测与规避仿真演示平台进行开发,解决了视景仿真中多视口显示、虚拟相机数据采集和阴影生成等关键技术,实现了火星软着陆末段障碍检测与规避技术的可视化仿真与演示。更多还原

【Abstract】 With the more deep exploration of Mars, the more complex terrain with higher scientific research value will go into detection range., which demands higher technique for soft landing task. The autonomous navigation technique with higher ability of obstacle detection and avoidance will become one of the key techniques in the future. In this paper, reference to the landing process of Phoenix Mars Lander, the research of the obstacle detection and hazard avoidance method is performed, and the related visual simulation platform is developed.First of all, based on the image sequence obtained by CCD camera, the crater and rock are detected and identified respectively. A novel contour detection method based on GVF-Snake dynamic model is given for the detection of crater, and a rapid detection methods based on multi-resolution analysis is presented for the detection of rock. By the means of the C-means clustering method, the target size and distribution of obstacles are calculated, and then, a safe landing site is obtained by spiral search.Secondly, according to the safe landing site, the research of hazard avoidance method is performed, a hazard avoidance method based on solving the special two-point boundary value problem and a algorithm based on Lyapunov stability method are given. The algorithm based on Lyapunov stability method take the threat of terrain dangerous into account, in this method, the Lyapunov function is designed as a compromise function of Energy function and terrain potential threat function,and then, the hazard avoidance guidance law is gained.Finally, the simulation and demonstration platform for obstacle detection and hazard avoidance is developed in the development environment of visual simulation engine Delta3D, the key techniques in visual simulation such as multi-view, camera data acquisition and shadow generation is implemented, At last, the visual simulation of the obstacle detection and hazard avoidance in the terminal descent of soft landing is achieved.更多还原