摘要
提出两种观点:1)可以使用一段惯性传感器数据完成捷联惯性导航系统(SINS)的初始对准;2)通过在数据融合中加入逆向—正向SINS解算结果和外部参考数据,可以缩短传感器误差的估计时间。基于以上两观点,旨在不改变卡尔曼滤波器的情况下,短时间内估计陀螺漂移的一种快速传递对准的方法。在一个参考数据更新周期内,储存了惯性传感器数据和参考数据,并且执行了逆向—正向捷联解算。与此同时,当相应的捷联解算结束后,执行数据融合算法。在上述的更新周期内,由于加入逆向—正向捷联解算,陀螺漂移估计操作增加了两次并且缩短了其估计时间。在舰船晃动的情况下,实验表明:与经典的方法相比,当两种方法的数据融合执行次数相等时,提出的方法缩短了对准时间,对准时间从300 s缩短到100 s;当对准时间相等时,方案1的纵摇、横摇和航向对准误差的均值分别为?0.6638、?0.5896、0.3941,其方差分别为1.0736、0.4629、2.1697,而方案2对应的均值和方差分别为?0.4632、?0.5026、0.3375和0.8828、0.3876、1.7289。由此可知,方案2的对准精度高于方案1。
Two viewpoints are presented： 1） we can use a set of inertial sensor data to complete the initial alignment of strapdown inertial navigation system（SINS）; 2） by adding the results of backward-forward SINS navigation solution and external reference data into data fusion, the estimation time of the sensor error can be shortened. Based on these viewpoints, a rapid transfer alignment method without changing the Kalman filter is introduced to estimate the gyro drift in a short time. In one reference data update cycle, the inertial sensor data and reference data are stored, and the backward-forward SINS solution is executed. Meanwhile, data fusion algorithms are performed after finishing the corresponding SINS solution. In the above update cycle, due to the addition of backward-forward SINS solution, the operation of gyro drift increases twice and its estimation time is shortened. The experiment on a shaking ship shows that the alignment time is shorten to 100 s by the proposed method, which is reduced by 3 times compared with that by the classical method, if the two methods takes equal time for data fusion. In situation that the two schemes take equal alignment time： the mean errors of pitch, yaw and heading alignment by scheme one are ？0.6638, ？0.5896, and 0.3941, respectively, and the variances are 1.0736, 0.4629, and 2.1697, respectively, while in scheme two, the mean errors are ？0.4632, ？0.5026, 0.3375, and the variances are 0.8828, 0.3876, 1.7289,showing that the alignment accuracy of the scheme two is higher than that of the scheme one.
出处
《中国惯性技术学报》
EI
CSCD
北大核心
2015年第6期727-732,共6页
Journal of Chinese Inertial Technology
基金
国家自然科学基金项目(61473085
51175082
51375088
61273056)
