以某型自行火炮炮载惯性导航系统为研究对象。为解决大方位失准角造成的系统非线性问题，在对大失准角误差模型进行详细分析的基础上，提出了基于快速正交搜索（FOS）和卡尔曼滤波（KF）的非线性参数估计方法。利用事先训练好的非线性误差模型进行对准，既能消除线性姿态误差，又可以对非线性姿态误差起到良好的抑制作用。仿真结果表明，FOS/KF方法的对准精度和实时性远优于扩展卡尔曼滤波（EKF）。对比试验结果表明，单独使用EKF时的方位角误差最大达到14.99°，而FOS/KF可以使方位角误差保持在0.8°以内。FOS/KF方法的估计精度不随系统非线性程度的变化而变化，并且不需要进行粗对准，简化了对准过程，提高了载体机动性。
The gun-board Strapdown Inertial Navigation System (SINS) is discussed. To solve the problem of system non-linear caused by the large azimuth misalignment angle, a non-linear parameter estimation method based on Fast Orthogonal Search (FOS) and Kalman Filter (KF) is proposed after a detailed analysis of the model for estimation of large misalignment angle errors. The non-linear error model previously trained is used for alignment. The linear attitude error can be eliminated, and the nonlinear attitude error can be suppressed effectively. The simulation results show that the FOS/KF method can result in more alignment accuraty and real-time than the Extended Kalman Filter (EKF). With a comparison of the test results shows that the azimuth angle error is up to 14.99° when the EKF is used alone, while the azimuth angle error is within 0.8° when the FOS/KF method is used. The accuracy of estimation with FOS/KF method does not change with nonlinearity of the system and does not need coarse alignment, simplifying the alignment process and improving the carrier maneuverability effectively.
[1] SAVAGE P G. Strapdown analytics[M]. Minn:Strapdown Associates, 2000:331-332.[2] GAIFFE T. From R&D brass board to navigation grade FOG-based INS:The experience of Photonetics/Ixsea[C]//Optical Fiber Sensors Conference Technical Digest. Piscataway, NJ:IEEE Press, 2012:1-4.[3] NAPOLITANO F, GAIFFE T, COTTREAU Y, et al. PHINS:The first high performances inertial navigation system based on fiber optic gyroscopes[C]//International Conference on Integrated Navigation Systems. Piscataway, NJ:IEEE Press, 2002:296-304.[4] EI-SHEIMY N, NASSAR S, NOURELDIN A. Wavelet de-noising for IMU alignment[J]. IEEE Aerospace and Electronic Systems Magazine, 2004, 19(10):32-39.[5] ARASA I, HAYKIN S. A numerical-integration perspective on Gaussian filters[J]. IEEE Trans on Automatic Control, 2009, 54(8):1254-1269.[6] WAN E A, VAN D, MERWE R. The unscented kalman filter for nonlinear estimation[C]//Adaptive System for Signal Processing Communication and Control, Piscataway, NJ:IEEE Press, 2000:1389-1392.[7] YU M J, PARK H W, JEON C B. Equivalent nonlinear error models of strapdown inertial navigation system[C]//ION GPS/GNSS 2003. Manassas, VA:ION Publications, 2003:188-198.[8] 李东明. 捷联式惯导系统初始对准方法研究[D]. 哈尔滨:哈尔滨工程大学, 2006:85-88. LI D M. Study on initial alignment methods of strapdown inertial navigation system[D]. Harbin:Harbin Engineering University, 2006:85-88(in Chinese).[9] 王志伟, 石志勇, 全振中. 惯性导航在线标定滤波技术[J]. 火力与指挥控制, 2014, 39(12):2061-2065. WANG Z W, SHI Z Y, QUAN Z Z. A survey of inertial navigation on-line calibration filtering technique[J]. Fire Control & Command Control, 2014, 39(12):2061-2065(in Chinese).[10] HASSIBI B, SAYED A H, KAILATH T. Linear estim-ation in Krein spaces-part2:Application[J]. IEEE Transactions on Automatic Control, 2016, 41(1):34-49.[11] WANG X Y, HUANG Y. Convergence study in extended Kalman filter-based training of recurrent neural networks[J]. IEEE Transactions on Neural Networks, 2016, 22(4):588-600.[12] GAN X S, DUANMU J S, CONG W. Wavelet neural network based on ssukf and its applications in aerodynamic force modeling for flight vehicle[C]//Proceeding of the 2000 National Technical Meeting of the Institute of Navigation. Anaheim, CA:Institute of Navigation, 2000:574-582.[13] 王巍. 惯性技术研究现状及发展趋势[J]. 自动化学报,2013, 39(6):723-728. WANG W. Status and development trend of inertial technology[J]. Acta Automatic Sinica, 2013, 39(6):723-728(in Chinese).[14] WU Y X, WU M P, HU X P, et al. Self-calibration for land navigation using inertial sensors and odometer:Observability analysis[C]//AIAA Conference of Guidance, Navigation and Control. Reston, VA:AIAA, 2013:1-10.[15] WU Y X, GOODALL C, EI-SHEIMY N. Self-calibration for IMU/Odometer land navigation:Simulation and test results[C]//Proceedings of ION/ITM. Manassas, VA:ION Publications, 2015:839-849.[16] KORENBERG M J. A Robust orthogonal algorithm for system identification and time series analysis[J]. Biological Cybernetics, 1989, 60(4):267-276.[17] KORENBERG M J, PAARMANN L D. Applications of fast orthogonal search:Time-series analysis and resolution of signals in noise[J]. Annals of Biomedical Engineering, 1989, 17(3):219-231.[18] KORENBERG M J. Identification nonlinear difference equation and functional expansion representations:The fast orthogonal algorithm[J]. Annals of Biomedical Engineering, 1988, 16(4):123-142.[19] KORENBERG M J, PAARMANN L D. Orthogonal approaches to time-series analysis and system identification[J]. IEEE Signal Processing Magazine, 1991, 8(3):29-43.[20] MIKAEL E J, KORENBERG M J, JAMES M P. Nonlinear system identification and control of chemical processes using fast orthogonal search[J]. Journal of Process Control, 2017, 17(9):742-754.[21] 高伟. 捷联惯性导航系统初始对准技术[M]. 北京:国防工业出版社, 2014:105-106. GAO W. Initial alignment for strapdown inertial navigation system[M]. Beijing:National Defense Industry Press, 2014:105-106(in Chinese).
CJA