Publications

Centralized Model Predictive Control for Multi-Agent Assembly of Space Structures

Abstract

In this paper, we develop a control architecture to address the challenge of maintaining control of a large space structure during on-orbit assembly. The goal is to coordinate the actions of discrete autonomous agents attached to the structure, each with their own onboard actuators. We discuss two approaches based on Model Predictive Control (MPC), in which the Optimal Control Problem (OCP) is repeatedly solved to generate optimal control trajectories over a receding time horizon. We present two approaches in this work - one in which a centralized MPC algorithm controls the entire network of agents, and one in which the MPC acts as an outer-loop planner while the agents use inner loop distributed trajectory tracking controllers to track the planned trajectory. Each method is tested for its ability to scale with the number of attached agents.

Date
2026
Authors
David Bacher, Avi Soval, Jackson D Alberto, Samarth Mankan, David Barnhart
Book
ASCEND 2026
Pages
3035