摘要
针对城市情况下车载导航时单一导航源易受干扰的问题,提出了一种基于自适应联邦Kalman滤波的多源组合导航算法。该模型具有两级结构,由子滤波器进行各信息源局部估计后,通过主滤波器进行最优融合估计。融合具有不同工作特点的导航传感器的输出信息组成多源信息组合导航系统,从而提高了导航系统的精度和鲁棒性,且通过故障诊断算法实时检测并隔离故障信息源。给出了联邦滤波算法设计,并进行了实际车载实验。实验结果表明,该算法能够提高导航系统的稳定性及精度。
A new adaptive multi-source fusion integrated navigation algorithm based on federated filtering is put forward for the problem of interference with single navigation source when driving in city.This model adopts a two-stage structure.After local estimation in the sub-filters,the optimal synthesis is performed in the main filter.The information of navigation sensors with different working characteristics is combined to form a multi-source information integrated navigation system,thereby improving the accuracy and robustness of the navigation system,and detecting and isolating the faulty sensors in real time through a fault isolation algorithm.The design of federated filtering algorithm is given and the actual vehicle experiment is performed.Experimental results show that the algorithm can improve the stability and accuracy of the navigation system.
作者
顾涛
陈帅
谭聚豪
王琮
陈安升
GU Tao;CHEN Shuai;TAN Ju-hao;WANG Cong;CHEN An-sheng(School of Automation, Nanjing University of Science and Technology, Nanjing 210094, China;Beijing Institute of Automation Control Equipment, Beijing 100074, China)
出处
《导航定位与授时》
CSCD
2021年第3期20-26,共7页
Navigation Positioning and Timing
关键词
多源融合
联邦滤波
故障隔离
组合导航
Multi-source fusion
Federated filtering
Fault isolation
Integrated navigation
