摘要
提出了基于全局采样的UPF算法,解决了UPF因其计算量大而难以应用于GPS/DR组合导航中的问题。新的算法通过结合最新的量测值,对粒子集整体利用一次UKF产生建议分布,省去了UPF中对每个粒子循环套用UKF的环节,减小了UPF的计算量。对状态变量为6维的GPS/DR组合导航非线性模型进行仿真实验,结果表明改进的UPF能够以较小的计算代价,实现对组合导航状态变量的精确估计。
The method of proposal distribution generation for unscented particle filter （UPF） based on global sampling is presented to solve the problem that it is difficult for UPF applied to the global posi- tioning system/dead reckoning（GPS/DR） integrated navigation because of its large computation. In order to reduce the computation of the UPF, it combines the latest measured values with used unscented Kal- man fiher（UKF） to generate a proposal distribution for the particle set, eliminating the need for UPF which applies UKF to each particle for circulation. Simulations use the observations in nonlinear model of GPS/DR integrated navigation. The results show that, improved UPF can get accurate estimation of state variables in integrated navigation at lower computational cost.
出处
《现代防御技术》
北大核心
2013年第5期47-52,共6页
Modern Defence Technology
关键词
UPF
建议分布
全局采样
GPS
DR组合导航
unscented particle filter （UPF）
proposal distribution
global sampling
global positioning system/dead reckoning（GPS/DR） combined navigation
