Robust approximate symmetric model predictive control

In this paper we study the explicit model predictive control (MPC) design for approximately symmetric systems. The approximately symmetric system is modeled as a symmetric system plus an additive disturbance residing in a symmetric set. The explicit MPC controller is computed using symmetry to minimize the number of critical regions stored in memory while being robust to model mismatch and disturbances. We show through numerical examples that approximating an approximately symmetric systems using a full symmetric system can drastically reduce memory usage and computation time for explicit controllers while still guaranteeing feasibility and input-to-state stability.