Model-Free Primitive-Based Iterative Learning Control Approach to Trajectory Tracking of MIMO Systems With...

Area of Science:

• Control Systems Engineering
• Robotics
• Artificial Intelligence

Background:

• Trajectory tracking in complex systems like MIMO requires precise control.
• Existing methods often rely on accurate system models, which are not always available.
• Iterative learning control (ILC) offers a framework for improving performance over repeated tasks.

Purpose of the Study:

• To develop a novel model-free trajectory tracking method for MIMO systems.
• To leverage primitives and ILC for efficient and adaptive control.
• To demonstrate the approach's capability in planning, reasoning, and learning.

Main Methods:

• A model-free iterative learning control (ILC) framework combined with primitives.
• Decomposition of complex trajectories into output primitives (basis functions).
• Optimization of reference input primitives without prior process knowledge, ensuring convergence via model-free virtual reference feedback tuning.

Main Results:

• Optimal trajectory tracking solutions derived from learned primitives.
• Efficient computation of optimal reference input without repeated task executions.
• Decomposition of MIMO optimization into decoupled single-input single-output problems for model-free decoupling.

Conclusions:

• The proposed model-free primitive-based ILC approach enables effective trajectory tracking for MIMO systems.
• The method demonstrates planning, reasoning, and learning capabilities.
• Validation through a case study on a nonlinear aerodynamic system confirms its efficacy.