Rotation Parameters Estimation of Space Target Based on Orbit Modeling and Levenberg-Marquardt Optimization

Inverse synthetic aperture radar (ISAR) imaging plays an important role in space target classification and recognition. For ISAR imaging of space target, the geometry model of space target orbit observation needs to be established before imaging, and then the expressions of radar line of sight vector (LOS), rotation vector and equivalent Doppler vector are obtained. Then the imaging coordinate system is established according to the accurate orbit prediction, and then the imaging is performed by using Range Doppler (RD) algorithm. If there is attitude adjustment in the operation of the space target, the rotation vector is composed of the rotation vector caused by the change of the radar line of sight vector and the rotation vector caused by the attitude adjustment of the target itself. Under the simplified assumption, if the target's attitude is adjusted to rotate at a constant speed around an axis in the system, and the rotation parameters are unknown, the actual synthetic rotation vector of the target cannot be obtained only by orbit prediction. In this paper, based on the orbit model of this kind of space target, the nonlinear equations are constructed according to the rotation speed estimates of multiple observation angles, and then transformed into a nonlinear least square problem. Finally, the optimal solution of self rotation parameters is obtained by using Levenberg Marquardt (LM) algorithm.