高斯粒子滤波的惯性／GPS紧组合算法

于永军¹, 徐锦法¹, 熊智², 张梁¹

(1. 直升机旋翼动力国家级重点实验室(南京航空航天大学), 210016 南京; 2. 南京航空航天大学 导航研究中心, 210016 南京)



摘要:

为提高组合导航系统的可靠性,针对以伪距、伪距率残差为量测信息的紧组合算法会带来线性化误差的缺点,推导了基于伪距、伪距率的非线性紧组合模型．针对紧组合系统状态维数高导致粒子滤波实时性差的问题,提出基于线性非线性结构分解的高斯粒子滤波算法,对状态方程中的非线性和线性部分利用高斯粒子滤波和经典卡尔曼滤波分别进行递推,有效降低了系统的运算量．仿真结果表明,使用改进的紧组合滤波算法系统定位精度相比线性化紧组合算法提高一倍．

关键词:  惯性导航  组合导航  紧组合  非线性滤波  高斯粒子滤波

DOI：10.11918/j.issn.0367-6234.2015.05.014

分类号:V249.3

基金项目:中国博士后科学基金(2013M541668)；江苏高校优势学科建设工程资助；江苏省博士后基金(1401041B).



SINS／GPS tightly integrated algorithm with gaussian particle filter

YU Yongjun¹, XU Jinfa¹, XIONG Zhi²,ZHANG Liang¹

(1.Science and Technology on Rotorcraft Aeromechanics Laboratory(Nanjing University of Aeronautics and Astronautics), 210016 Nanjing, China; 2. Navigation Research Center, Nanjing University of Aeronautics and Astronautics, 210016 Nanjing, China)

Abstract:

To improve the reliability of integrated navigation, a tightly coupling nonlinear model based on pseudo range and rate is proposed for reducing the estimation error of tradi-tional algorithm using linear measurement model in this work. For the application of particle filtering to SINS/GPS tightly integrated navigation system, the dimension of the state variables has been a major constraint for the Real-time system. In this new arithmetic, a linear KF deduction and nonlinear GPF method have been employed for the linear part and the non-linear part to improve the precision and real time performance, respectively. Results from the simulation show that the hybrid algorithm can effectively improve the performance of the integrated navigation system, and the precision increases one time.

Key words:  inertial navigation  integrated navigation  tightly coupling  non-linear filter  gaussian particle filter