基于人工神经网络的地形线性化方法

Terrain Linearization Method Based on Neural Network

中文关键词  人工神经网络  地形线性化  扩展卡尔曼滤波  地形辅助导航系统

英文关键词  neural network  terrain linearization  extended Kalman filtering (EKF）  terrain aided navigation (TAN)

中文摘要

      地形线性化方法对地形辅助导航系统性能有重要影响，本文采用人工神经网络这一新技术来分块记忆地形数据，给出数字地图上任意位置的线性化参数，并用于系统状态的扩展卡尔曼滤波，与其他传统地形线性化方法相比，用人工神经网络来处理地形函数，系统滤波性能得到改善，提高了导航精度，并可减少地图的存贮量，适于并行处理。

英文摘要

      The terrain linearization method has significant influence on terrain aided navigation system (TANS). In this paper artificial neural network is applied to fit terrain data and calculate the parameters of linearization at any position on the digital map, then these parameters used in extended Kalman filtering. The simulation results show that the navigation accuracy of TANS can be raised by using neural network terrain linearization method. This method can save memory space of digital map and fits paralled calculation.