Robust-Switched Model Predictive Control for a Class of Hybrid Rotors/Reaction-Wheels Driven In-Cabin Space Robots

This article focuses on the switched system modeling and model predictive control (MPC) for a class of hybrid rotors and reaction-wheels driven in-cabin space robots, which is capable of switching the enabled actuators to reach balance between high maneuverability and low interference. With randomly changed payloads and operation objectives, the space robots are modeled by Markov jump systems with mode-dependent constraints. Different from traditional Markov jump systems, mode switching between different actuation/payload modes is considered to be separated by practical dwelling stages, such that the sojourn time is lower-bounded. Accordingly, the switched MPC approach is extended to allow for stochastic performance optimization with lower-bounded Markov switching in the prediction horizon. Also, a class of bounded disturbances is considered, and the proposed robust MPC ensures the stability and satisfaction of hard constraints by incorporating robust positive invariant sets and disturbance-weighted cost. Compared with previous studies, the issues of constrained lower-bounded Markov jump system modeling and robust-switched MPC design for hybrid-actuation space robots are addressed for the first time. The experimental results are provided to demonstrate the effectiveness and merits of the proposed approach.