基于3星子集的GPS快速选星算法  

GPS Fast Constellation Selection Based on 3-SAT Subset

作　　者：霍国平[1,2] 缪玲娟[1] 高志峰[1] 

机构地区：[1]北京理工大学自动化学院,北京100081 [2]清华大学信息技术研究院,北京100081

出　　处：《宇航学报》2014年第5期574-580,共7页Journal of Astronautics

基　　金：国家自然科学基金(61153002)

摘　　要：针对高精度GPS导航系统中,空间星座数量变化时由星座选择带来的运算量较大的问题,利用Sherman-Morrison矩阵求逆引理,推导得到GDOP(Geometric Dilution of Precision)值的增量递推计算公式。在此基础上提出一种基于由3颗GPS卫星组成"3星子集"的快速选星算法,并利用LLRB树(Left Leaning Red-Black Tree)的存储搜索策略辅助快速产生最佳4星组合。相对于传统GDOP选星法,在可视星卫星数增加时,浮点数运算量(FLOPs)可减少将近一半;当可视卫星数减少时,FLOPs可降低到接近为0。实际试验结果表明,3星子集选星方法可以有效降低星座突变时由星座选择带来的时间消耗,提高星座更新的实时性。A fast constellation selection method named ＇ 3-SAT subset ＇ is presented for reducing the large amount of computation when a sudden change occurs in the sky view in high-accuracy GPS navigation system. The GDOP （ Geometric Dilution of Precision） recursive formula is deduced by use of the Sherman-Morrison formula to reduce matrix inversion operation. Moreover, efficient storage and search strategy of LLRB-Tree （Left Leaning Red-Black Tree） algorithm is used to assist rapid creation of optimal 4-SAT combinations and get the minimal GDOP. With respect to the traditional GDOP constellation selection, the amount of floating-point operations per second （FLOPs） can be reduced by more than half when the number of visualization satellites increases, and almost equals 0 when the number decreases. The experimental results show that, 3-SAT subset method can effectively reduce the time consumption at the mutation of constellations and improve the real-time performance accordingly.

关 键 词：星座选择 几何精度因子 LLRB树 浮点数运算量 

分 类 号：TN967.1[电子电信—信号与信息处理]