FFRLS-MPC: An adaptive model predictive control method for robust ship course keeping.

Area of Science:

• Maritime Navigation and Control Systems
• Adaptive Control Theory
• Ocean Engineering

Background:

• Accurate course control is critical for maritime safety and efficiency.
• Conventional controllers struggle with dynamic vessel changes and environmental forces.
• Existing methods may not adequately address model-plant mismatch and unmodelled disturbances.

Purpose of the Study:

• To develop an adaptive course-control strategy for enhanced maritime navigation.
• To improve controller performance under time-varying dynamics and environmental disturbances.
• To provide a robust solution for high-precision ship motion control.

Main Methods:

• Coupling online system identification with Model Predictive Control (MPC).
• Utilizing Fast Forgetting Recursive Least Squares (FFRLS) for real-time parameter identification.
• Integrating an adaptive model into MPC for state prediction and rudder optimization.

Main Results:

• The FFRLS-MPC controller demonstrated faster responses and lower steady-state errors compared to PID and robust adaptive MPC.
• Accurate dynamic course tracking was achieved under various simulated environmental conditions.
• The identification module effectively adapted to disturbance effects through parameter variation.

Conclusions:

• The proposed FFRLS-MPC strategy offers a practical approach to robust, high-precision ship motion control.
• Real-time system identification significantly enhances controller adaptability and performance.
• This method effectively compensates for model-plant mismatch and unmodelled environmental disturbances.