球面单形平方根无迹粒子滤波在拖曳合成孔径声纳组合导航中的应用

Spherical simplex square-root unscented particle filter used in integrated navigation system of synthetic aperture sonar

惯性导航设备修正回波数据作为合成孔径声纳(SAS)运动补偿的一种重要方法,其中的关键技术是研究和利用高精度的滤波算法以及组合导航系统为运动补偿提供导航数据信息。在标准粒子滤波算法的基础上进行改进,提出了一种球面单形平方根无迹粒子滤波算法(SS-SRUPF),并将之应用于SINS、多普勒计程仪(DVL)组成的水下组合导航系统,输出的导航数据再传输给SAS成像系统。仿真结果表明,SS-SRUPF算法可将一个合成孔径时间内的组合导航位置均方误差精度提高到0.013m,小于SAS距离向分辨力0.019 m,满足SAS运动补偿的要求。

In motion compensation of synthetic aperture sonar(SAS), one of important methods is using inertial navigation system to correct the echo data. In this paper, a high-precision filter algorithm and integrated navigation system were studied and utilized for providing navigation data to realize the motion compensation. A new algorithm, i.e. spherical simplex square root unscented particle filter(SS-SRUPF), is proposed and applied into the integrated SINS/DVL navigation system. The navigation data from the integrated system is then transferred to the SAS imaging system. Simulation results indicate that the integrated navigation precision in a synthetic aperture time is increased to 0.013 m(RMS) which is smaller than the SAS range differentiability of 0.019 m, and the computation efficiency is more than twice the UPF’s, meeting the motion compensation requirement of SAS.