地形特征对粒子滤波导航算法精度影响分析

Influence of topographic feature on particle filter terrain

摘要:

地形导航是利用飞行器下方地形估计飞行器空间位置的导航方法。由于地形是强非线性模型，因此地形导航可以看作非线性状态估计问题，地形特征直接影响导航精度。在讨论粒子滤波导航算法基础上，重点研究了地形特征对后验概率密度和Cramer-Rao下界的影响，得出了地形梯度和相似度是影响粒子滤波导航算法精度的根本原因，为飞行区域选择、飞行路径规划提供了依据，具有较强的实用性。

Abstract:

Terrain navigation is the method that can estimate the position of aircraft from the terrain information of the area below the aircraft. As terrain height is a strongly nonlinear model, terrain navigation can be solved as a nonlinear state estimation problem, and the accuracy of its algorithm is directly influenced by the topographic feature. After deducing the particle filter navigation algorithm, an emphasized study is conducted about the influence of topographic feature on the calculation of posterior probability, and Cramer-Rao Lower Bound(CRLB). The conclusion is that the gradient and comparability in topographic feature are the essential characters to cause effects on the result of particle filter algorithm, which can provide reference for the flight zone selection and the flight route planning.