State and Parameter Estimation in Space Vehicles With Robotic Manipulators on SE(3) and TSE(3)

This paper explores the dynamic complexities of robotic manipulators (RMs) in space missions. The challenges posed by the RMs' movements, which alter the mass properties of the spacecraft and potentially lead to instability and unpredictable dynamics, are analyzed. To address these challenges, the paper details the application of an on-manifold unscented Kalman filter (UKF) estimation scheme to predict and adjust for changes in mass parameters along with the states of the vehicle, enhancing the accuracy of future estimation-based control algorithms. Dealing with a chain of rigid bodies in the vehicle, this work emphasizes the use of a joint square-root UKF on the tangent bundle of the special Euclidean group (TSE(3)) to simultaneously estimate the states and mass parameters, replacing previous dual methodologies for application to systems with time-variant mass properties. This avoids the system excitation requirements inherent to the previous dual methodology. A square-root filter is incorporated to ensure the symmetry and positive definiteness of the state covariance matrix is maintained through the algorithm. Numerical simulations demonstrate the effectiveness of the estimation scheme for a spacecraft-manipulator system in a low lunar orbit.