因子图发展及其在定位与导航的应用技术

因子图作为一种表示因式分解的建模工具,在编码领域、统计学、信号处理和人工智能领域有着广泛的应用．因子图在导航领域的应用研究逐步发展起来．与单一导航系统对比,组合导航系统能够提供更精确、更具鲁棒性的导航结果,但是因其各个子系统的误差特性与工作频率不同的特点,增加了导航系统的设计复杂性．基于因子图的组合导航算法可以有效解决导航信息融合中的传感器异步问题且实现对多传感器的灵活配置,使得系统具有即插即用的特性,在非线性量测条件下可以获得较好效果.导航系统中的状态估计以及信息融合问题可以使用因子图模型表示,基于因子图的和 积算法是组合导航信息融合的主要算法.本文对因子图及其在导航系统中的应用进行了探讨,主要包括:1)因子图的数学理论基础及其相关应用领域;2)因子图在定位与导航领域的发展和应用. 

Development of factor graph and its application technology in positioning and navigation

Factor graphs are widely used in the fields of coding,statistics,signal processing,and artificial intelligence as a modeling tool that represents factorization.The application of factor graphs in the navigation field is also gradually developed.The combination of multiple sensor information provides a more accurate and robust navigation state estimate than the information from a single sensor.However,various sensors have different error characteristics and these sensors usually operate at different frequencies.Considering that some sensors cannot supply the linear measurement information,and it brings the challenges to the design of the integrated navigation system.The navigation algorithm based on the factor graph model enables the system to have plug-and-play characteristics and achieve better results under nonlinear measurement conditions.The state estimation and information fusion problems in the navigation system can be represented by the factor graph model.Sum-product algorithm based on factor graph is the main method in navigation and positioning system.This paper summarizes the factor graph theory and its application in the navigation system,including:1)Mathematical theoretical basis of factor graph and its related application fields 2)Development and application of factor graphs in the field of positioning and navigation.