摘要
针对故障检测阈值难以确定和系统非线性较强时精度差的问题,提出了1种新的多传感器容错融合算法。各量测子系统采用无迹信息滤波,不需要对系统进行线性化,因而能避免线性化误差。设计基于模糊逻辑的故障检测算法。利用卡方检验值计算子系统有效概率。根据子系统的有效概率自适应隔离故障子系统,处理故障检测阈值难以选取的问题。将该文算法应用于捷联惯性导航系统/北斗2/全球定位系统(SINS/BD2/GPS)组合导航系统中,仿真结果验证了其有效性。
A new multi-sensor fault-tolerant fusion algorithm is proposed for the poor accuracy of nonlinear data fusion systems and fault detection threshold selection problem. Unscented information filters are embedded into measurement subsystems, and the linearization error is avoided because the system need not be linearized. A fault detection algorithm based on fuzzy logic is designed. The subsystem validation probability is calculated using Chisquare test. Fault subsystems are isolated adaptively according to the validation probability of the subsystem, and the difficulty of selecting fault detection threshold is solved. The proposed algorithm is applied to a strap-down inertial navigation system/beidou2/global position system （ SINS/BD2/GPS ） integrated navigation system, and the simulation results show the availability.
出处
《南京理工大学学报》
EI
CAS
CSCD
北大核心
2017年第3期269-277,共9页
Journal of Nanjing University of Science and Technology
基金
国家自然科学基金(61673219
61673217)
关键词
无迹信息滤波
多传感器
容错
数据融合
模糊逻辑
卡方检验
unscented information filter
multi- sensor
fault tolerant
data fusion
fuzzy logic
Chisquare test
