An adaptive fading kalman filter algorithm for rtk

Abstract

The invention belongs to the field of satellite navigation differential receiver design, and relates to a design algorithm for solving dynamic adaptability of a floating-point solution of ambiguity by using an adaptive fading Kalman filter. The invention adds the single-difference integer ambiguity into the state vector of the filter model for filter estimation, and introduces the forgetting factor to improve the dynamic tracking performance of the filter. In addition, an independent sliding window is introduced to collect and process innovations, which solves the problem that the covariance of innovations cannot be accurately calculated due to interruption of observations due to changes in reference stars or loss of satellite signals. The application of the algorithm is simple, there is no complex or huge calculation formula, and the real-time and calculation can guarantee the realization of the algorithm programming on the DSP or FPGA hardware platform, that is, it is easy to implement in engineering.

Description

technical field [0001] The invention belongs to the field of satellite navigation differential receiver design, and relates to a design algorithm for solving dynamic adaptability of a floating-point solution of ambiguity by using an adaptive fading Kalman filter. Background technique [0002] Carrier phase-based dynamic differential technology (RTK technology) is a high-precision positioning technology, which has a wide range of applications, such as high-precision navigation guidance, UAV approach and landing, and so on. The key of RTK technology is the fast and accurate solution of the ambiguity of the whole circle. Among them, the most widely used ambiguity resolution algorithm is the LAMBDA algorithm, which is divided into two processes: ambiguity real number estimation and ambiguity integer search. The real estimate of the ambiguity provides an initial search value for the ambiguity parameter, usually a floating point number. The high-precision ambiguity floating-poin...

AI Technical Summary

Problems solved by technology

However, when the Kalman filter is used for real-time dynamic differential positioning, the motion state of the maneuvering carrier is complex and changeable, and the system model of the filter is difficult to establish accurately. Due to the poor robustness of the standard Kalman filter, it is easy to cause divergence

At present, the first-order time correlation model (Singer model) and the "current" statistical model are generally used in the dynamic filtering model. They are all improvements to the system model. Compared with the constant velocity model and the constant acceleration model, the dynamic performance of the filter is somewhat different. improved, but still did not fundamentally solve the problem

Images