A relative pose estimation method for non-cooperative targets in space based on upf

Abstract

The invention discloses a method for estimating the relative pose of a space non-cooperative target based on UPF, which includes the following steps: initialization of the relative pose and three-dimensional coordinates of SURF feature points, tracking and matching of SURF feature points, and relative pose estimation based on UPF . The present invention introduces the Bayesian filter method and measurement inversion method used for pose estimation in the field of visual real-time positioning and map construction SLAM to the field of spatial vision navigation, and can measure the relative pose of a completely non-cooperative target, which can be applied In-orbit services, such as the capture of space debris, can also be applied to space offense and defense, to obtain status information of non-cooperative targets, and to serve for further identification and operational decision-making. The present invention can provide high-precision pose output, has good anti-noise performance, does not need prior information such as the shape and motion state of the target, is not limited to assumptions about the target state, and has good autonomy, versatility and convenience sex.

Description

technical field [0001] The invention relates to the technical field of spatial visual navigation, in particular to a method for estimating relative pose and attitude of non-cooperative targets in space based on unscented particle filtering UPF. Background technique [0002] The development of space technology has increased the need to complete complex tasks such as capture or transfer of space debris and abandoned satellites, repair or replacement of faulty in-orbit spacecraft, satellites with extended life through refueling, etc., which require tracking / missions The spacecraft accurately estimates the relative position and attitude of the target at close range. In the non-cooperative state, the motion and structure of the target are completely unknown or partially unknown, and there is no inter-satellite link to assist in the pose measurement, so the relative pose measurement of the non-cooperative target is challenging and difficult. According to whether the geometric mod...

AI Technical Summary

Problems solved by technology

However, the above two methods assume that the initial pose of the target is known, which is not conducive to practical application

Schnitzer et al. of Dresden University of Technology aimed at the problem of autonomous maintenance of space-failed spacecraft, and proposed an algorithm based on EKF-SLAM and RANSAC to estimate the relative pose of the target and restore the three-dimensional structure of the target, but this method relies on the extended Kalman filter EKF For the linearization of nonlinear equations, when the system is highly nonlinear, excessive linearization errors will lead to a decrease in filtering performance

Images