基于卡尔曼滤波器的在轨目标跟踪算法研究

发布时间：2018-01-04 23:26

本文关键词：基于卡尔曼滤波器的在轨目标跟踪算法研究　出处：《中国科学院大学(中国科学院光电技术研究所)》2017年硕士论文　论文类型：学位论文

【摘要】：随着人类对太空探索的不断深入,太空逐渐成为了世界大国角力的主战场。对空间的控制和利用已经成为一个国家科学技术水平的象征。对太空中大量在轨目标进行准确且稳定的跟踪是对其进行观测或控制的前提条件。通常使用轨道根数合理外推出在轨目标的位置矢量和速度矢量等状态信息,再依据状态量对在轨目标的观测信息进行解算,并以此引导观测设备进行跟踪。工程应用中,由于轨道根数会随着时间发生蠕变或轨道根数测不准等原因,使轨道外推得到的在轨目标状态量存在一定误差,该误差会导致目标不在视场中心,从而产生脱靶量。经纬仪视场很小,脱靶量的存在会影响跟踪效果,甚至导致丢失跟踪目标。针对该问题,本文重点讨论了针对非合作目标在仅测角条件下的滤波跟踪问题。首先,研究了描述在轨目标运动的时间尺度以及天球坐标系和地球坐标系的转换,并使用SOFA软件库实现了坐标转换算法;其次,提出了一种基于复杂力学模型的滤波算法,该力学模型中考虑到了在轨目标在空间中运动所受到的主要摄动力因素,并对一些摄动力模型进行了简要的阐述,在此基础之上使用数值积分对摄动力方程进行求解,实现了轨道外推算法,将该算法与STK软件中的高精度轨道预报函数的计算结果进行了比较;之后,算法以在轨目标的位置矢量和状态矢量为系统状态量,以目标相对于地基观测站(经纬仪)的高度角和方位角为观测量设计了基于扩展卡尔曼滤波算法和无迹卡尔曼滤波算法的滤波器。最后,使用MATLAB和C搭建了算法的仿真测试平台,以STK软件和部分实测数据为算法输入,研究了不同帧频角度数据和滤波器参数对滤波结果的影响。仿真实验结果表明,本文所研究的在轨目标滤波跟踪系统中,扩展卡尔曼滤波算法与无迹卡尔曼滤波算法都可以通过多组经纬仪测角数据来修正含有误差的轨道根数,无迹卡尔曼滤波算法在鲁棒性、收敛速度和滤波精度上略优于扩展卡尔曼滤波算法,但一定程度上牺牲了算法的运算效率。
[Abstract]:With the deepening of space exploration, space has gradually become the main battlefield of world power. Control of space and use has become a symbol of national science and technology level. The accurate and stable tracking of space in a large number of objects in orbit is a prerequisite for measurement or control of the position vector and. The velocity vector of state information usually use orbital launch reasonable orbital target, according to the state of objects in orbit observation information is solved, and guide observation equipment tracking. In engineering application, because the orbit will occur as time creep or orbit uncertainty and other reasons, there are some errors in the target state of the track extrapolated, the error will lead to the target is not in the center of the visual field, resulting in the miss distance. The theodolite field is very small, there will be the shadow of miss distance Sound tracking effect, and even lead to loss of tracking targets. Aiming at this problem, this paper mainly discusses the conditions in bearings only tracking filter for non cooperative target problems. First, change of describing objects in orbit movement time scale and celestial coordinate system and earth coordinate system, and use SOFA software to realize the coordinate conversion secondly, put forward the algorithm; a filtering algorithm based on complex mechanical model, mainly considering the on orbit perturbation target motion in space by the dynamic factors of the mechanical model, and some perturbation force model are discussed, on the basis of using the numerical integral solution of the perturbation equation, the orbit extrapolation algorithm, the calculation function of the high precision orbit prediction algorithm and STK software are compared with the results of; after the algorithm takes the position vector and state vector of the target in orbit As the system state variables, relative to the ground observation station to target (theodolite) the altitude angle and the azimuth angle measurement is designed for the concept of extended Calman filter algorithm and unscented filter algorithm based on Calman filter. Finally, using MATLAB and C to build a simulation test platform of the algorithm, using STK software and part of the measured data for algorithm input of different frame angle data and filter parameters on the filtering results. Simulation results show that the orbital target filter tracking system introduced in this paper, the extended Calman filter algorithm and unscented filter algorithm can Calman through multiple sets of theodolite data containing orbit error, no trace of Calman filtering algorithm in robustness, convergence speed and accuracy is slightly better than that of the extended Calman filter algorithm, but to a certain extent at the expense of the operation efficiency of the algorithm.

【学位授予单位】：中国科学院大学(中国科学院光电技术研究所)
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TN713

【相似文献】