大失准角情形下UKF与CKF的比较研究

Comparative Study on UKF and CKF in Large Azimuth Misalignment for SINS

扼要介绍了UKF和CKF滤波算法,分析总结了两种滤波算法的异同,并基于蒙特卡罗仿真试验,对大方位失准角情形下UKF滤波算法和CKF滤波算法的应用特性进行了对比、分析和总结,指出了UKF滤波算法对角运动和线运动均敏感,角运动和线运动方向的激励均可以加速其收敛,而CKF滤波算法对角运动不敏感,但对线运动敏感,角运动方向的激励无助于其对准精度的提高,而线运动方向的激励可加速其收敛,且收敛精度随激励效果的加强而提高的特性,为该种情形下对准技术的工程化实现提供了针对性建议。

The Unscented Kalman Filter （UKF） and Cubature Kalman Filter （CKF） are introduced, and an analysis is made to their similarities and differences. Through Monte-Carlo simulation tests, the characteristics of UKF and CKF under large azimuth misalignment are studied, and the following conclusions are obtained ： 1 ） UKF is sensitive to both angular motion and linear motion, while CKF is only sensitive to linear motion;2） Both the incentive of angular motion and linear motion can speed up the convergence of UKF algorithm, but as to the CKF, the incentive of angular motion makes no contribution to the alignment accuracy while the incentive of linear motion can speed up the algorithm convergence;and 3 ） The CKF algorithm accuracy can be gradually increased with the improvement of the incentive of linear motion. The unique characteristics are helpful for the engineering realization of alignment technology.