测绘学报

• 学术论文 • 上一篇    下一篇

GPS/Doppler导航随机模型的移动窗口实时估计算法

周泽波1,沈云中2,李博峰3   

  1. 1. 电子科技大学
    2. 同济大学测量系现代工程测量国家测绘局重点实验室
    3. 同济大学测量与国土信息工程系
  • 收稿日期:2010-03-02 修回日期:2010-06-22 出版日期:2011-04-25 发布日期:2011-04-25
  • 通讯作者: 周泽波

Moving Time-window Based Real-time Estimation Algorithm for the Stochastic Model of GPS/Doppler Navigation

  • Received:2010-03-02 Revised:2010-06-22 Online:2011-04-25 Published:2011-04-25

摘要: GPS/Doppler随机模型对导航解算结果有重要影响,动态条件下需要实时估计随观测条件而变化的随机模型。本文引进了时间遗忘因子和观测冗余度因子,有效地平衡了移动窗口内不同时刻的观测数据及其冗余情况对单位权方差估值的贡献,改进了单位权方差的移动开窗实时估计算法;并采用载噪比模型确定观测权阵,等价权抗差估计方法处理粗差。实测车载GPS/Doppler数据的处理结果表明:采用本文算法显著提高了GPS/Doppler的导航精度。

Abstract: The stochastic model of GPS/Doppler play an important role in the navigation solution. Two-step method for estimating the stochastic model of GPS/Doppler integrated observation is proposed. Firstly, the robustly posterior estimation is introduced according to equivalent weight principle and carrier-to-noise ratio, Then, the variance of unit weight is estimated in a moving -window in real time by combining the forgetting factor and observation redundancy. Finally, a real vehicular GPS data is carried out to test the proposed method.