International Conference on Instrumentation, Measurement, Circuits and Systems (ICIMCS 2011)
116 The Resampled Kernel-Diffeomorphism Filter For GPS/INS Integration
Download citation file:
- Ris (Zotero)
- Reference Manager
Because the Global Positioning System (GPS) and the Inertial Navigation System (INS) complement each other, it is common practice to integrate the two systems in applications calling for continuous high accuracy and reliability. In this paper we propose a resampled kernel-diffeomorphism filter (RKDF) as a variation of particle filter with improved sampling efficiency and performance for GPS/INS integration. The RKDF invokes kernel-diffeomorphism to form a continuous estimate of the posterior density function. The advantage of the kernel-diffeomrphism estimator (KDE) is the suppression of the border convergence difficulties by using an appropriate regular change of variable. Simulation results demonstrate that the RKDF gives better estimate than the standard extended Kaiman filter (EKF).