Orbit Determination of an Uncooperative RSO using a Stereo Vision-Based Sensor

Space Situational Awareness (SSA) involves the capability of identification, tracking and characterization of resident space objects. This study considers the use of an inspector spacecraft to autonomously determine the orbit and ballistic coefficient of a resident space object (RSO) in real-time using a stereo vision- based relative navigation sensor when in close proximity with the RSO. Additionally, the inspector spacecraft performs a leader-follower and circumnavigation type formations with the RSO to facilitate 3-D reconstruction of its shape from the visual images. An Unscented Kalman Filter (UKF) is formulated to process raw measurements in the two-dimensional pixel coordinate frame from the visual images of the RSO. The UKF produces relative orbit estimates of the RSO with respect to the inspector in addition to its ballistic coefficient. The UKF uses a novel formulation of relative dynamics that includes differential drag effects and applied thrust accelerations by the inspector spacecraft. The inspector\u27s own orbit and ballistic coefficient are estimated by processing raw GPS pseudorange measurements using an Extended Kalman Filter. The inspector\u27s orbit estimates can be fused with the relative orbit estimates of the RSO to autonomously determine the orbital elements of the RSO. A linear optimal controller is formulated to enable the inspector spacecraft to maintain the RSO in field-of-view of the two cameras as well as to perform leader-follower and circumnavigation type formations. Simulation results showing the feasibility of using stereo vision-based sensors for meeting the above stated SSA objectives and proximity operations are presented