Model Predictive Control With Guaranteed Feasibility of Inequality Path Constraints.

Area of Science:

• Control Engineering
• Optimization Theory
• Applied Mathematics

Background:

• Existing nonlinear model predictive control (MPC) methods struggle to guarantee path constraints (PCs) feasibility.
• Direct multiple shooting enforces PCs only at discrete time points, risking constraint violation.

Purpose of the Study:

• To develop a novel nonlinear MPC framework ensuring both stability control and guaranteed feasibility of PCs.
• To address the computational burden associated with guaranteeing PCs feasibility in MPC.

Main Methods:

• Application of semi-infinite programming to MPC's rolling optimization.
• Introduction of an event-triggered sampling mechanism to reduce computational complexity.
• Derivation of sufficient conditions for asymptotic convergence of closed-loop systems.

Main Results:

• A novel MPC framework guaranteeing path constraint feasibility during rolling optimization.
• An event-triggered algorithm effectively reducing computation while maintaining PCs feasibility.
• Demonstrated effectiveness on a cart-damper-spring system.

Conclusions:

• The proposed event-triggered MPC framework ensures stability and path constraint satisfaction.
• This approach offers a computationally efficient solution for path-constrained control problems.
• The method is validated through simulation on a relevant mechanical system.