A novel parameterization method to estimate the relative state and inertia parameters for non-cooperative targets

This paper presents a novel parameterization method to estimate the relative pose, motion and inertia ratios of a non-cooperative target using stereoscopic vision only. As there is no a prior information about the structure, shape and inertia parameters of the target, the traditional model matching method is unavailable. Besides, the angular velocity of the target cannot be measured directly without gyro embedded on the target. By parameterizing the inertia parameters in a new method, the Euler's rotational dynamics is used to propagate the angular velocity of a free-floating non-cooperative target. Two cameras are mounted on the chaser spacecraft to acquire images of the target and then the 3D positions of some feature points are obtained. Based on the stereo vision measurements, an Extended Kalman Filter is developed to estimate the relative state, including the relative position, relative attitude, relative linear velocity, and relative angular velocity, as well as the inertia ratios. The main contribution of this paper is that the inertia ratios are parameterized by two dependent variables using hyperbolic tangent functions, which fully considers the physical constraints between the inertia parameters. Numerical simulations are carried out to evaluate the validity of the proposed algorithm.