PPP/INS tightly coupled navigation using adaptive federated filter
详细信息 查看全文
- 作者：Zengke Li ; Jingxiang Gao ; Jian Wang ; Yifei Yao
- 关键词：PPP ; INS ; Adaptive ; Federated filter ; Tightly coupled navigation
- 刊名：GPS Solutions
- 出版年：2017
- 出版时间：January 2017
- 年：2017
- 卷：21
- 期：1
- 页码：137-148
- 全文大小：
- 刊物类别：Earth and Environmental Science
- 刊物主题：Geophysics/Geodesy; Atmospheric Sciences; Space Sciences (including Extraterrestrial Physics, Space Exploration and Astronautics); Automotive Engineering; Electrical Engineering;
- 出版者：Springer Berlin Heidelberg
- ISSN：1521-1886
- 卷排序：21
文摘
Integration of the global positioning system (GPS) with inertial navigation system (INS) has been very intensively studied and widely applied in recent years. Conventional GPS/INS integrated systems that receive pseudorange and Doppler observations can only attain meter-scale accuracy. An INS has also been integrated with double-differenced GPS measurements that remove GPS errors, although this increases system cost. Following the availability of real-time precise orbit and clock products, a precise point positioning PPP/INS tightly coupled navigation system is presented here. Because various types of measurements such as pseudorange, carrier phase and Doppler are available, an adaptive federated filter method is proposed and applied to the PPP/INS integrated system to improve filter efficiency and adaptivity. Provided that the federated local filter and the adaptive filter are equivalent in form, an information allocation factor in the federated filter is constructed based on the adaptive filter factor. Simulation analyses for different INS grades show that the tactical grade INS can provide higher initial value accuracy for PPP. An experiment was performed to validate the new algorithm, and the results indicate that the INS can improve PPP accuracy, especially under challenging positioning conditions. PPP solution accuracy in the east, north and down components can improve by 45, 47 and 24 %, respectively, during partial GPS satellite blockages. The resolution accuracy of the proposed adaptive federated filter is similar to that of a centralized Kalman filter. The proposed method can also realize parallel filter computing and remove the influence of dynamic model errors.