中国空间科学技术

中国空间科学技术 ›› 2020, Vol. 40 ›› Issue (4): 61-68.doi: 10.16708/j.cnki.1000-758X.2020.0046

• 研究探讨 • 上一篇    下一篇

多传感器组合导航系统的改进多尺度滤波算法

林雪原，孙玉梅，董云云，乔玉新，陈祥光   

  1. 1 烟台南山学院电气与电子工程系，烟台265713
    2 北京理工大学，北京100081
  • 出版日期:2020-08-25 发布日期:2020-07-20

Improved multiscale filtering algorithm of multisensor integrated navigation system#br#

LIN Xueyuan, SUN Yumei, DONG Yunyun, QIAO Yuxin, CHEN Xiangguang   

  1. 1 Department of Electrical and Electronic Engineering，Yantai Nanshan University，Yantai265713，China
    2 Beijing Institute of Technology，Beijing100081，China

  • Published:2020-08-25 Online:2020-07-20

PDF (PC)

310

摘要： 多尺度滤波算法在多传感器组合导航系统中已得到成功应用，然而该算法用到多个时刻的量测向量，导致算法计算量过大，并影响系统的实时性。针对上述问题，首先利用分块技术与小波变换将时域内描述的系统原始状态方程转换为块状态方程，然后将实时得到的当前时刻的量测向量表达为块状态向量的形式，最后结合常规卡尔曼滤波技术与序贯滤波的思想，提出了一种改进的多传感器组合导航系统多尺度滤波方法。将该算法应用于GPS/SST/SINS多传感器组合导航系统，仿真结果验证了该算法不仅具有较好的实时性，而且相对于传统算法,系统的定位精度提高1倍以上。

关键词: 多尺度滤波, 改进算法, 卡尔曼滤波, 组合导航系统, 块状态方程

Abstract: Multi-scale filtering algorithm has been successfully used in multisensor integrated navigation system, but this algorithm uses measurement vectors at different times, thus causing heavy computation burden and affecting the system′s real-time performance. According to the shortcomings of the application, the block partition technique and wavelet transformation were used to change the system state equation described in timedomain into block state equation. Then,the real-time measurement vector was expressed in the form of block state equation. At last, an improved multi-scale filtering method of multi-sensor integrated navigation system was established using Kalman filter and sequential filter. The algorithm was applied to the GPS/SST/SINS multisensor integrated navigation system, and the simulation results show that this algorithm has not only better realtime and recursive performance, but also better fused precision.

Key words: multi-scale filter, improved algorithm, Kalman filter, integrated navigation system, block state equation