Title: GPS Receiver Failure Detection Method in High Dynamic Environment
Author(s): Junyi Li, Yandong Wang
Published in: Proceedings of IEEE/ION PLANS 2018
April 23 - 26, 2018
Hyatt Regency Hotel
Monterey, CA
Pages: 349 - 354
Cite this article: Li, Junyi, Wang, Yandong, "GPS Receiver Failure Detection Method in High Dynamic Environment," Proceedings of IEEE/ION PLANS 2018, Monterey, CA, April 2018, pp. 349-354.
Full Paper: ION Members/Non-Members: 1 Download Credit
Sign In
Abstract: In high dynamic environment, the accuracy of GPS positioning is polluted by multiple failures. The failure types of GPS receiver in high dynamic environment are investigated in the paper, such as divergent position error and surge of velocity error, which are caused by rapid attitude changing and high dynamic stress on tracking loop. To detect the GPS failures, it is common to use the Receiver Autonomous Integrity Monitor (RAIM). However, the application of this method is constrained by some conditions. In the paper, Geometric Dilution of Precision (GDOP) check method, Chi-Square residual method and double-state Chi-Square method are presented for detecting failures in high dynamic environment. These approaches combine the innovation and raw observations to promote failure detection. As the rocket changes attitude rapidly, the satellites in view vary fast, therefore the coefficient matrix of single point positioning, which affects GDOP, reconstructs simultaneously. In order to make sure the position accuracy meets the requirements, GDOP is used as the detected value. Meanwhile, the auxiliary of Strap-Down Inertial Navigation (SINS) is utilized for failure detection. To verify the effectiveness of the proposed methods, a self-developed simulation platform covering all procedures of failure detection is employed. The key challenge of these methods is to set the appropriate thresholds for different detected values. For GDOP, the threshold can be derived from the estimated variances of position errors, while others can be calculated based on the variance of the detected value and the false alarm rate. As the result of the simulation, the performance of each method under different simulation environments is evaluated and the best method is selected for the specific failure.