神经网络辅助的GPS/INS组合导航故障检测算法  

Neural Network Aided GPS/INS Integrated Navigation Fault Detection Algorithms

作　　者：高为广[1,2] 杨元喜[3] 

机构地区：[1]信息工程大学测绘学院 [2]61081部队 [3]西安测绘研究所

出　　处：《测绘学报》2008年第4期403-409,共7页Acta Geodaetica et Cartographica Sinica

基　　金：国家自然科学基金(40604003);“863”项目(2007AA12Z331)

摘　　要：针对GPS/INS松组合导航系统观测信息无冗余,而且观测信息可能存在故障的情形,提出一种神经网络辅助的组合导航故障检测算法。该算法克服了基于模型的故障检测算法受模型误差影响的局限性;能够自动地对观测信息进行故障的检测、定位和剔除;能够基于故障检测后可靠的观测信息进一步调整动力学模型信息对导航解的贡献;能够在GPS失锁时,较好地进行导航预报。最后利用车载实测数据进行验证,结果表明该算法能够很好地从模型误差中分离出观测信息含有的故障信息,降低了故障检测算法存在的虚警率,避免故障信息对导航解的影响;且GPS失锁时,神经网络的预报输出在一定程度上能够进一步提高导航解的精度。Aiming at the characteristic of lacking observation and possible faults of loosely-coupled GPS/INS inte grated navigation, a neural network aided integrated navigation fault detection algorithm is put forward. The new algorithm is helpful to localize the measurement outliers when the kinematic model has significant errors, and can automatically detect and isolate the faults in the case that there are not redundant observations, and can balance the contributions of the dynamical model information and reliable measurements on the state vector estimates, and especially can reasonably predict the navigation results when GPS outages occured. It is shown, by comparison and analysis, that the new algorithms can not only separate the faults from the dynamical model er ror, reduce the false alarm rate and avoid the influences of faults on the navigation results, but also improve the accuracy of navigation solutions during GPS outages.

关 键 词：GPS/INS组合导航 神经网络 故障检测 自适应估计 自适应滤波 

分 类 号：P228.4[天文地球—大地测量学与测量工程]