In-Space Demonstration of Model Predictive Control Approaches for Space Towing of Uncertain Loads

In this paper, model predictive control (MPC) approaches are implemented to deal with spacecraft maneuvering problems for cargo carrying with an uncertain load for future in-space servicing and assembly missions. First, a tube-based robust MPC approach is tested using Astrobee robots onboard the International Space Station (ISS). Its performance and capability are demonstrated when an active Astrobee moves an inactive Astrobee as an unknown object to a desired position. Second, a decentralized MPC framework is exploited to transport an unknown object to a desired location using two active Astrobees. Based on a leader-follower strategy using a virtual leader and formation constraints, the two robots can maintain a specific formation and achieve multi-agent cooperative transportation tasks. Simulation and experimental results are compared to demonstrate the decentralized MPC approach's capability.