A dual-rate hybrid filtering method to eliminate high-order position errors of GPS in POS

The Position and Orientation System (POS) serves as a key component for the airborne remote sensing system, which integrates Strapdown Inertial Navigation System (SINS) and Global Position System (GPS) to provide the reliable and continuous motion compensation using Kalman Filter (KF). However, the high-order position errors resulting from C/A (Coarse/Acquisition) Code GPS cannot be effectively compensated or estimated by the traditional KF, which severely weakens the imaging quality. In this paper, we propose a Dual-rate Hybrid Filter (DHF) to deal with the high-order position errors based on Least Squares Support Vector Machine (LSSVM) and Kalman Filter. DHF builds a low update rate filter by integrating high-precision SINS and online LSSVM to isolate the high-order position errors. Meanwhile, the high update rate filter of DHF maintains the advantages of traditional SINS/GPS integrated navigation system to restrain the accumulation errors of system. The experimental results show that the proposed method significantly reduces the high-order position errors by 84.6% at each sampling period comparing with the conventional single KF based SINS/GPS integrated navigation system.