On-Orbit Mass Property Estimation of Combined Spacecraft with Sparse Identification of Nonlinear Dynamics

We present a new approach for on-orbit identification of mass properties of combined spacecraft systems following docking events. Assuming known chaser spacecraft properties, we identify the mass, center of mass, and principal moments of inertia of the combined system after docking with an unknown object. The method uses a gray-box approach combining physics-based equations of motion with data-driven techniques: nonlinear regression for mass and center of mass identification, and Sparse Identification of Nonlinear Dynamics (SINDy) for inertia estimation. Unlike black-box models that only predict behavior, SINDy recovers interpretable equations describing the physical motion, enabling direct comparison with known governing equations to extract mass properties. To address gyroscope measurement noise sensitivity, we employ Weak SINDy (WSINDy), an integral formulation that attenuates high frequency noise through integration rather than differentiation. Our numerical simulations confirm the accuracy and robustness of the methods to measurement noise. The approach enables on-orbit adaptation to changing system configurations, offering a fast, data-efficient solution for autonomous spacecraft operations.