自适应CDPF及其在组合导航中的应用

Adaptive CDPF for Integrated Navigation

在组合导航系统中,高精度的导航算法对导航解算精度有非常重要的影响。为提高捷联惯导(SINS)/合成孔径雷达(SAR)组合导航系统定位的解算精度,针对粒子滤波中密度分布函数难以选取的问题,提出一种新的自适应中心差分粒子滤波(CDPF)算法。通过中心差分卡尔曼滤波来获取状态均值和协方差阵,计算获得自适应因子,并利用得到的因子自适应的调节均值和方差信息,得到一种参数可调节的重要性密度分布函数,提高了滤波精度,使滤波性能明显改善,能更好地解决非线性非高斯系统模型的滤波问题。将提出的算法应用于SINS/SAR组合导航系统中,仿真结果表明,改进的滤波算法能提高导航定位的解算精度,系统优化性能明显优于扩展卡尔曼滤波、粒子滤波以及中心差分粒子滤波。

In the integrated navigation system, high precision navigation algorithms have an important influence on navigation calculation precision. In order to improve the navigation positioning accuracy of the strapdown inertial navi- gation system（ SINS）/synthetic aperture radar（SAR） integrated navigation systems and aim at the particle filtering problem that it is difficult to select the importance density function, this paper presented an adaptive central difference particle filtering （CDPF） algorithm by adopting central difference Kalman filtering to obtain state estimation and co- variance, and then the adaptive factor calculated can adaptively regulate the estimation and covariance information. It provides adjustable parameters of importance density function and is more suitable for filtering calculation based on nonlinear and non-Gaussian models, through considering the latest measurement information and improving the parti- cle filtering performance. The proposed algorithm was applied to SINS/SAR integrated navigation system. Simulation results demonstrate that the adaptive CDPF outperforms the extended Kalman fihering, particle filtering and CDPF in terms of positioning calculation accuracy in navigation system.