摘要
针对车载捷联惯导系统工作环境恶劣且子惯导系统精度低,使用传统的Kalman滤波进行传递对准存在精度低、实时性差的问题。根据车载特殊环境提出了一种基于LS-SVM的大方位失准角速度匹配传递对准方法,采用传统Kalman滤波器的输入输出样本对LS-SVM滤波器进行训练,避开了经典SVM求解时的优化运算,使其复杂度大大降低。在相同条件下与传统Kalman滤波仿真结果对比,LS-SVM方法的引入可以使系统对准时间缩短近50%,并适当提高了系统的对准精度。
Aiming at harsh working environment of on board strapdown inertial navigation system (SINS) and problem of low precision of sub-inertial navigation system and poor alignment precision and real-time in transfer alignment of conventional Kalman filtering. Transfer alignment method for big azimuth misalignment angle speed matching based on LS-SVM is presented according to on-board special environment. Using input and output sample pair of traditional Kalman fiher to train the LS-SVM filter, and avoid complex optimization computing in classical support vector machine (SVM)solving. Under the same conditions, compare with the conventional Kalman filtering simulation result, introduction of LS-SVM method can make system alignment time shorten nearly 50 %, and increase system alignment precision appropriately.
出处
《传感器与微系统》
CSCD
北大核心
2013年第1期145-148,152,共5页
Transducer and Microsystem Technologies
基金
国家自然科学基金资助项目(61031001
60972118)
北京市创新人才项目(PHR201006115)
科技创新平台项目(71F1210907)
"国家十二五"预先研究专题项目(40405100304)