Design of a robust model predictive controller with reduced computational complexity.

Area of Science:

• Control Engineering
• Computational Mathematics
• Systems Science

Background:

• Robust Model Predictive Control (MPC) is crucial for systems with uncertainties.
• Computational time for solving optimization problems limits MPC practicality.
• Existing methods often face challenges with computational complexity.

Purpose of the Study:

• To reduce the computational complexity of robust model predictive control.
• To present a novel method for decreasing the time consumed in solving the robust MPC optimization problem.
• To maintain control performance while enhancing computational efficiency.

Main Methods:

• A scaled state vector is defined to simplify objective function contours.
• The control input is determined via state feedback minimizing an infinite horizon objective function.
• Linear matrix inequalities are solved to find the optimal control input.

Main Results:

• The number of iterations required to solve the optimization problem at each sampling interval is reduced.
• Control performance remains largely unchanged compared to conventional methods.
• The proposed method demonstrates significant computational savings.

Conclusions:

• The developed method effectively reduces computational complexity in robust MPC.
• This approach enhances the practicality of robust MPC for real-time applications.
• The technique offers a viable solution for time-sensitive control systems.