Singular Value Decomposition H∞ Federal Filtering and Application in Transfer Alignment

GAO She-sheng,JIANG Wei-wei,SONG Fei-biao

DOI: https://doi.org/10.13695/j.cnki.12-1222/o3.2011.05.003

2011-01-01

Abstract:In strapdown inertial navigation system(SINS)/synthetic aperture radar(SAR) integrated navigation systems,the anti-interference and dynamic rapid transfer alignment between SINS and the additional inertial measurement unit(IMU) of the antenna for SAR is a difficult problem.In order improve the speed and precision of transfer alignment without increasing the order of filter,and reduce the amount of calculation,the paper presents a new singular value decomposition H_∞ federal filtering algorithm by absorbing the advantages of existed filtering methods.First,the model of transfer alignment for velocity + attitude + position matching update is established,and the robust filtering is carried on by the slave filter for H_∞ filtering which is based on singular value decomposition to obtain the local optimum estimation of system states.Then,the global optimum estimation is derived based on the local optimum estimation fused by the federal filter.The simulation results demonstrate that the misalignment angles' estimations,which are obtained by the proposed filtering algorithm,are converged on 3′,-5′and 20′in 200 s respectively,and the filtering performance is significantly superior to that of H_∞ filtering and federal filtering.The new filtering algorithm has the characteristics of fast speed,high accuracy,small calculation amount and good anti-interference performance,providing an effective approach to improve the accuracy of transfer alignment.