Dynamic Event-Triggered MPC With Shrinking Prediction Horizon and Without Terminal Constraint

Area of Science:

• Control Systems Engineering
• Automation and Robotics
• Dynamic Systems Analysis

Background:

• Model Predictive Control (MPC) traditionally relies on terminal constraints for stability guarantees.
• Event-triggered and self-triggered control offer potential for reduced computational load but can introduce conservatism.
• Existing methods often struggle to balance triggering frequency with system performance.

Purpose of the Study:

• To develop a dynamic event-triggered model predictive control (MPC) strategy.
• To eliminate the need for stability-related terminal constraints in MPC.
• To improve the efficiency of control systems by optimizing triggering and measurement frequency.

Main Methods:

• A novel dynamic event-triggering mechanism is proposed, integrating aspects of both event- and self-triggering.
• The prediction horizon is dynamically adjusted, shrinking as system states converge.
• Stability is proven without relying on traditional terminal constraints, and recursive feasibility of the optimization control problem (OCP) is guaranteed.

Main Results:

• The proposed dynamic event-triggered MPC strategy successfully stabilizes the system.
• Conservatism in triggering rate and measurement frequency is explicitly addressed and reduced.
• The shrinking prediction horizon contributes to system convergence and stability.

Conclusions:

• The developed dynamic event-triggered MPC offers a viable alternative to traditional MPC with terminal constraints.
• This approach enhances control system efficiency and performance.
• Simulation results validate the effectiveness of the proposed stabilization strategy.