Robust Multiplexed Model Predictive Control for agent-based conflict resolution

Multiplexed Model Predictive Control (MMPC) was originally developed for a multi-input system as a strategy to reduce online computation. In this paper, we demonstrate how distributed control of a system of agents can be posed as a Multiplexed Model Predictive Control problem. Specifically, we consider using robust multiplexed MPC for controlling a system of agents in the presence of coupling constraints in the form of a collision avoidance requirement. The system is subject to persistent unknown, but bounded disturbances. The contribution of this paper is the extension of the original robust multiplexed MPC algorithm to include a disturbance feedback policy in between updates. This facilitates finding feasible solutions and inherits the property of rapid response to disturbances from multiplexing the control. In addition, it is observed that the computational time of the proposed MMPC scheme scales favourably with the number of agents.