Nonlinear MPC Design using the qLPV Approach and IQC-based Terminal Ingredients

Area of Science:

• Control Engineering
• Systems Theory
• Nonlinear Systems

Background:

• Model predictive control (MPC) for nonlinear systems often uses sector arguments, which can be conservative.
• Integral quadratic constraints (IQCs) with dynamic multipliers offer less conservative designs and larger regions of attraction.
• Nonlinear MPC can be computationally demanding due to nonlinear prediction models.

Purpose of the Study:

• To develop a systematic nonlinear MPC design procedure.
• To combine IQC-based terminal ingredients with linear parameter varying (LPV) tools.
• To reduce the computational burden of nonlinear MPC while maintaining performance.

Main Methods:

• Utilized IQCs with general dynamic multipliers for terminal ingredient synthesis.
• Integrated linear parameter varying (LPV) tools to address computational complexity.
• Developed a systematic nonlinear MPC design procedure combining these approaches.

Main Results:

• The proposed MPC scheme demonstrates good control performance.
• The method achieves a lower associated numerical cost compared to state-of-the-art nonlinear MPC.
• A processor-in-the-loop experiment validated the scheme's practical applicability.

Conclusions:

• The combined IQC and LPV approach offers a less conservative and computationally efficient nonlinear MPC design.
• This method enhances stability regions and reduces computational load.
• The validated scheme is suitable for practical engineering applications.