A dual-frequency combined filter and filtering method for eliminating single-point GPS high-order position errors

Abstract

The invention relates to a dual-frequency combination filter and a filtering method for eliminating single-point GPS high-order position errors, including an input part, a high-frequency combination part, a low-frequency combination part and an output part; the input part includes a strapdown inertial measurement unit SIMU, GPS module; the high-frequency combination part includes strapdown solution A and Kalman filter KF; the low-frequency combination part includes strapdown solution B and LSSVM filter; the output part includes a motion compensation unit; the structure of the high-frequency combination part is the same as that of the traditional SINS / The structure of GPS integrated navigation is similar. The low-frequency combination part isolates single-point GPS high-order position errors through the independence of strapdown solution B, and the high-order errors in the position information output by high-frequency combination are filtered out by the LSSVM online learning and training model. . In addition, at the end of each low-frequency filtering cycle, the predicted value of LSSVM provides position calibration for strapdown solution B, which prevents its position accumulation error from entering the real-time motion compensation unit, and properly handles the position information between the various low-frequency segments Articulation and compensation issues.

Description

technical field [0001] The present invention is a kind of dual-frequency combined filter and filtering method based on Least squares support vector machine (Least squares support vectormachine, LSSVM) to eliminate single-point GPS (Global Position System, Global Positioning System) high-order position error, can be applied to POS (Position and Orientation System, POS), specifically related to aerial remote sensing synthetic aperture radar (Synthetic Aperture Radar, SAR) high-resolution real-time imaging motion compensation. Background technique [0002] High-precision real-time motion imaging is the core of high-resolution aerial remote sensing systems. It requires the flight platform to perform ideal motion, but the flight platform will form complex multi-mode non-ideal motion under the influence of external environment and internal disturbances, resulting in blurred and scattered motion images. Focused, deformed, or even unable to image. Therefore, in order to achieve hig...

AI Technical Summary

Problems solved by technology

[0007] The problem solved by the technology of the present invention is: aiming at the problem that the traditional SINS / GPS integrated navigation method cannot meet the index requirements of high-order position errors of aerial remote sensing high-resolution real-time imaging, a dual-point GPS high-order position error elimination method is proposed. Frequency combination filter and filtering method, online elimination of the impact of GPS high-order position error on combined measurement output

Images