相邻历元误差相关的抗差卡尔曼滤波算法分析
作者简介:王仁(1990-),男,硕士生,主要从事GNSS数据处理理论及应用研究。E-mail: wangr1990322@163.com
收稿日期: 2015-03-30
要求修回日期: 2015-07-02
网络出版日期: 2015-12-20
基金资助
国家自然科学基金项目(41174032)
江苏省自然科学基金项目(BK20150236)
江苏师范大学研究生科研创新计划重点项目(2015YZD004)
The Analysis of Adjacent Epoch Error Related Robust Kalman Filtering Algorithm
Received date: 2015-03-30
Request revised date: 2015-07-02
Online published: 2015-12-20
Copyright
王仁 , 赵长胜 , 张敏 , 孙鹏 , 杜希建 . 相邻历元误差相关的抗差卡尔曼滤波算法分析[J]. 地球信息科学学报, 2015 , 17(12) : 1506 -1510 . DOI: 10.3724/SP.J.1047.2015.01506
According to the Kalman filtering theory and robust theory, we derived the model of adjacent epoch error related robust Kalman filtering algorithm. This model has a good robustness for observations containing gross errors. Through the analysis of deformation monitoring data containing gross errors and compare it to the model of adjacent epochs error related Kalman filtering algorithm, it can be concluded that using the proposed robust Kalman filtering model in data processing, regardless of whether or not there are gross errors in the observation values, the results of deformation analysis are consistent with the actual situation, which is not sensitive to the impact of gross error. And during the analysis of deformation monitoring data, we found that when the Kalman filtering method is used to estimate the state vector, it does not store a large amount of historical data, but takes use of the new observational data, through the continuous prediction and correction to estimate the new state of the system.
Tab. 1 Deformation monitoring data and the processing results of Kalman filtering algorithm表1 变形监测点数据及卡尔曼滤波处理结果(mm) |
期数 | 1 | 2 | 3 | 4 | 5 |
---|---|---|---|---|---|
观测值 | 92.690 | 93.160 | 92.830 | 92.390 | 92.210 |
方案一 | 92.808 | 93.135 | 92.887 | 92.424 | 92.199 |
方案二 | 92.808 | 93.135 | 92.887 | 92.424 | 92.199 |
方案一差值 | 0.118 | -0.025 | 0.057 | 0.034 | -0.011 |
方案二差值 | 0.118 | -0.025 | 0.057 | 0.034 | -0.011 |
期数 | 6 | 7 | 8 | 9 | 10 |
观测值 | 92.150 | 91.860 | 91.280 | 90.910 | 90.340 |
方案一 | 92.135 | 91.869 | 91.303 | 90.908 | 91.294 |
方案二 | 92.135 | 91.869 | 91.303 | 90.908 | 90.349 |
方案一差值 | -0.015 | 0.009 | 0.023 | -0.002 | 0.954 |
方案二差值 | -0.015 | 0.009 | 0.023 | -0.002 | 0.009 |
期数 | 11 | 12 | 13 | 14 | 15 |
观测值 | 89.690 | 89.090 | 88.760 | 88.540 | 88.300 |
方案一 | 89.777 | 89.082 | 88.739 | 88.524 | 88.294 |
方案二 | 89.699 | 89.091 | 88.747 | 88.528 | 88.295 |
方案一差值 | 0.087 | -0.008 | -0.021 | -0.016 | -0.006 |
方案二差值 | 0.009 | 0.001 | -0.013 | -0.012 | -0.005 |
期数 | 16 | 17 | 18 | 19 | 20 |
观测值 | 88.180 | 87.960 | 87.850 | 87.490 | 87.070 |
方案一 | 88.172 | 87.960 | 87.846 | 87.498 | 87.557 |
方案二 | 88.173 | 87.961 | 87.846 | 87.498 | 87.077 |
方案一差值 | -0.008 | 0 | -0.004 | 0.008 | 0.487 |
方案二差值 | -0.007 | 0.001 | -0.004 | 0.008 | 0.007 |
期数 | 21 | 22 | 23 | 24 | 25 |
观测值 | 86.380 | 85.780 | 85.060 | 84.610 | 84.270 |
方案一 | 86.422 | 85.781 | 85.063 | 84.602 | 84.262 |
方案二 | 86.394 | 85.784 | 85.066 | 84.603 | 84.262 |
方案一差值 | 0.042 | 0.001 | 0.003 | -0.008 | -0.008 |
方案二差值 | 0.014 | 0.004 | 0.006 | -0.007 | -0.008 |
期数 | 26 | 27 | 28 | ||
观测值 | 83.710 | 83.210 | 83.070 | ||
方案一 | 83.713 | 83.210 | 83.058 | ||
方案二 | 83.714 | 83.210 | 83.058 | ||
方案一差值 | 0.003 | 0 | -0.012 | ||
方案二差值 | 0.004 | 0 | -0.012 |
The authors have declared that no competing interests exist.
[1] |
|
[2] |
|
[3] |
|
[4] |
|
[5] |
|
[6] |
|
[7] |
|
[8] |
|
[9] |
|
[10] |
|
[11] |
|
[12] |
|
[13] |
|
[14] |
|
[15] |
|
[16] |
|
[17] |
|
[18] |
|
[19] |
|
[20] |
|
[21] |
|
[22] |
|
[23] |
|
[24] |
|
[25] |
|
[26] |
|
[27] |
|
[28] |
|
/
〈 | 〉 |