Deep-space exploration navigation system filtering method with immunity to initial error

Abstract

The invention provides a nonlinear filtering method with immunity to initial error, for solving the problem of filtering of a deep-space exploration high-precision autonomous navigation system. The method uses a parallel filtering structure, and comprises the following steps of: generating an initial value of a sub-filter of each channel through UT (Unscented Transformation), then, for each channel, obtaining the state and the variance estimation value of each sub-filter according to a computing method of EKF (Extended Kalman Filtering), finally, calculating a weighting coefficient of an autoregressive method, and weighting the previous calculation results to obtain the final state and variance estimation value. The method of the invention can avoid influence on filtering precision caused by inaccurate distribution of the initial values, namely, the method has certain immunity to the initial error, so that the method is very suitable for the deep-space exploration type navigation system of which the initial value obtaining precision is not high.

Description

technical field [0001] The invention belongs to the field of spacecraft navigation, guidance and control, and relates to a filtering method for deep space exploration and navigation systems. Background technique [0002] With the in-depth study of Kalman filter theory, various nonlinear filter algorithms have emerged, among which the extended Kalman filter (EKF) and the unscented Kalman filter (UKF) are widely used. The convergence and precision of these filtering algorithms have a certain dependence on the selection of the initial value, and the initial value of the filter obeys the Gaussian distribution, which means that the selected initial value may deviate far from the true value. The specific situation is as follows: figure 1 As shown, the circular curve represents a 3δ ellipse, the center of the circular curve is the coordinate origin, which represents the true value x(0) of the initial value, and the dense points in the center represent the estimated value of the ini...

AI Technical Summary

Problems solved by technology

[0003] The autonomous navigation system for deep space exploration has high requirements on the accuracy of the filtering algorithm, and the selection of the initial value of the system is relatively difficult, that is, the initial error of navigation is relatively large

If the existing EKF or UKF filtering algorithm is selected, the filtering accuracy depends on the selection of the initial value, so it is impossible to obtain higher navigation accuracy

Images