Data Fusion Approach for Altitude Location Error Estimation for Unmanned Aerial Vehicle Using Federated Filter

Yong-Hong Hu,Zhong-Ke Shi,Wei Tang
DOI: https://doi.org/10.1109/icit.2005.1600731
2005-01-01
Abstract:Altitude location for UAV by using federated filter is discussed, the fourth structure is selected, because its two sub-filters involving altitude sensor and the difference Global Positioning System (d-GPS) respectively are fully isolated from each other and thus make the federated filter more fault-tolerant and better in real-time performance. Data fusion based on this federated filter was simulated. When the d-GPS is working normally and simulation results show that values estimated by data fusion base on federated filter are very close to real values and the variance of federated filter converges to 10 m. When the electromagnetic environment is very bad, as is usually the case in war zone, the d-GPS is silent for a long time, variance based on federated filter converges to 13 m, bigger than the 10 m. Compared with previous paper variance based on Kalman converges to 15 m, higher than the 13 m. Federated filter can give much more accurate estimation than Kalman filter.
What problem does this paper attempt to address?