降阶扩展卡尔曼滤波算法在船用INS/GPS组合导航系统中的应用

摘要

TNNS(真航向导航系统)由MS860接收机、INS及处理数据的PC/104架构的嵌入式工控机构成.针对TNNS推导了INS(惯性导航系统)的误差模型,提出了适合于TNNS的降阶扩展卡尔曼滤波算法组合GPS和INS.系统在东海作了三次海试,软件及滤波算法平台由C/C++编制.海上试验表明,组合滤波后,INS的位置误差由100 m降低到40 m以下;进行最优化滤波后的航向误差σ由原来的0.105°减小为0.034°,纵横摇的误差也大幅减小.整个海试结果表明,在TNNS中组合GPS/INS采用的降阶扩展卡尔曼滤波算法,大幅提高了系统精度和可靠性.%TNNS (True North Navigation System) consist of MS860 GPS, INS(inertial navigation system), and a PC/104 embedded computer system for processing data. An INS error model was developed for the actual TNNS.Based on the error model, a reduced-order EKF(Extended Kalman Filter) algorithm was designed for TNNS. The system tests have been conducted three times at East China Sea. TNNS software and the reduced-order EKF algorithm are programmed by C/C++ codes, and executed under real-time mode in the embedded computer. The test results show that the INS position errors in north and east are reduced from more than 100 m to 40 m after correction, and the heading error σ is reduced from 0.105° to 0.034° after optimal filter correction. The accuracy of roll and pitch is also increased greatly by the integration of MS860 GPS/INS. The whole tests validate the effect of the algorithm, and show that the TNNS performance is significantly improved by the reduced-order EKF.