Static Output-Feedback H_∞ Control Design Procedures for Continuous-Time Systems With Different Levels of Model...

Area of Science:

• Control Systems Engineering
• Robotics
• Automation

Background:

• Optimal control design for continuous-time linear systems is crucial for performance.
• H-infinity control criteria offer robustness guarantees.
• Output-feedback control is desirable for practical implementation.

Purpose of the Study:

• To develop model-based and model-free optimal solutions for H-infinity control.
• To design static output-feedback controllers with flexible exponential criteria.
• To provide algorithms for both offline and online control design.

Main Methods:

• Formulation of a general H-infinity control design criterion with an exponential term.
• Development of two offline policy-iteration algorithms.
• Extension to a family of online off-policy designs.
• Adaptation for partial or complete model knowledge.

Main Results:

• A collection of model-based and model-free optimal output-feedback controllers are proposed.
• Algorithms accommodate varying levels of model information.
• Partial model knowledge reduces online matrix computations.
• Off-policy learning is feasible without disturbance excitation when the disturbance input matrix is known.

Conclusions:

• The presented methods offer versatile design alternatives for H-infinity control.
• The approach is validated through an optimal lane tracking controller for automated vehicles.
• The work facilitates robust control design in scenarios with limited or unavailable disturbance measurements.