An Adaptive Dynamic Kalman Filtering Algorithm Based on Cumulative Sums of Residuals
In order to overcome the drawbacks of the fault detection method based on \( \chi^{2} \) test that is insensitive to soft fault detection, an adaptive dynamic robust Kalman based on variance inflation model was developed, which can detect the soft fault o
- PDF / 1,968,891 Bytes
- 9 Pages / 439.37 x 666.142 pts Page_size
- 4 Downloads / 164 Views
An Adaptive Dynamic Kalman Filtering Algorithm Based on Cumulative Sums of Residuals Long Zhao and Hongyu Yan
Abstract In order to overcome the drawbacks of the fault detection method based on v2 test that is insensitive to soft fault detection, an adaptive dynamic robust Kalman based on variance inflation model was developed, which can detect the soft fault of system. The proposed method cumulates the residuals in open windows. When the cumulant surpasses the threshold, the error covariance is enlarged to prevent abnormal Global Positioning System (GPS) observations. This method has been applied to integrated navigation system of Inertial Navigation System/ Global Navigation Satellite System (INS/GNSS). The simulation results show that the soft fault is detected by using adaptive dynamic robust Kalman, and the filtering precision is higher than the traditional Kalman filtering algorithm. Keywords Integrated navigation filtering
Fault detection Robust filtering Kalman
67.1 Introduction Integrated navigation technology is an effective way to improve the overall performance of the navigation system. Inertial navigation system (INS) is autonomous, anti-interference, hidden, reliable, real-time and accurate in the short term. However, the accuracy decreases as the time is evolving. Hence, the INS and GPS satellite navigation systems are combined into INS/GPS integrated navigation system, which can complement each other, and increase the redundancy. Thus, L. Zhao (&) Science and Technology on Aircraft Control Laboratory, Beihang University, Beijing, China e-mail: [email protected] H. Yan Digital Navigation Center, Beihang University, Beijing, China
J. Sun et al. (eds.), China Satellite Navigation Conference (CSNC) 2013 Proceedings, Lecture Notes in Electrical Engineering 245, DOI: 10.1007/978-3-642-37407-4_67, Ó Springer-Verlag Berlin Heidelberg 2013
727
728
L. Zhao and H. Yan
multidimensional navigation information of highly data update rate can be obtained, and estimation accuracy and reliability of the navigation system can also be improved. INS/GPS integrated navigation system has been widely applied to autonomous navigation and positioning system of aircraft [1, 2]. With the higher requirements of high accuracy and reliability for integrated navigation system, the higher demand is presented to the fault detection and isolation of navigation system. The fault of the system can be divided into three types: hard fault, abrupt fault and soft fault. The hard fault can be detected through build-in test for navigation system; abrupt fault can be detected by the traditional fault detection methods. However, the soft fault of navigation system are difficult to detect, such as the filtering divergence caused by the uncertainty of the system model, the uncertainty of positioning caused by the GPS signal is disturbed, et al. In order to improve the reliability and accuracy of navigation system, dynamic navigation positioning was completed after detecting and isolating wrong GPS observation [3]. Many fault det
Data Loading...
