Model Predictive Control for Small Spacecraft Rendezvous with Non-cooperative Maneuvering Targets

This paper presents a two-phase model predictive control (MPC) framework for autonomous small spacecraft rendezvous with noncooperative maneuvering targets. The proposed approach employs distinct control periods and cost function weights for long-range orbit maneuvers and short-range rendezvous. To handle unknown target maneuvers, the framework integrates an adaptive filtering strategy that switches between unbiased minimum-variance input and state estimation (UMVISE) and extended Kalman filter. Simulations using a 3U CubeSat model validate the proposed control architecture. Simulation results demonstrate that the controller successfully guides a 3U CubeSat from initial separations of 925km to final rendezvous configurations. The UMVISE-MPC integrated tracking system maintains position tracking errors below 5cm during target maneuvers, enabling a successful rendezvous under the condition of noncooperative target behavior.