Attitude-orbit Integration Parameter Estimation Method for Space Non-cooperative Targets Based on Deep Learning

Abstract

The invention discloses a method for estimating parameters of attitude-orbit integration of space non-cooperative targets based on deep learning. The parameter estimation algorithm uses the dual vector quaternion to model the kinematics and dynamics of the space non-cooperative target, and on this basis, the corresponding state BP neural network parameter estimation algorithm and the state covariance matrix convolutional neural network parameter estimation are designed algorithm. The entire parameter estimation algorithm utilizes the characteristics of the dual vector quaternion to estimate the attitude-orbit parameters of the space non-cooperative target, and considers the attitude-orbit coupling effect of the space non-cooperative target. At the same time, this parameter estimation algorithm designs a BP neural network with a single hidden layer and a convolutional neural network with a double hidden layer, which can estimate the parameters of the non-cooperative target in the space under the condition of measurement failure, so that the parameter estimation algorithm is effective for the space environment. Strong robustness.

Description

technical field [0001] The invention belongs to the field of aerospace technology, and in particular relates to a method for estimating parameters of attitude-orbit integration of space non-cooperative targets based on deep learning. Background technique [0002] The ever-increasing amount of space debris has seriously affected normal human spaceflight activities. In particular, the growing number of failed spacecraft not only occupies a large amount of orbital resources, but also poses a great threat to space security. In recent years, major aerospace countries and international research institutions have reached a general consensus: in order to ensure the availability of orbital resources and the safety of space, it is necessary to remove space debris in space, especially failed spacecraft. During this process, due to failure, the spacecraft cannot provide its own information and rolls freely in space. Therefore, accurate parameter estimation is very important for subseq...

AI Technical Summary

Problems solved by technology

Under measurement failure conditions, if these parameters are estimated using traditional neural networks, it is prone to overfitting

Images