Autonomous berthing path tracking of a 4-DOF ship under nonlinear model predictive control

Area of Science:

• Marine engineering
• Robotics
• Control systems

Background:

• Unmanned surface ships (USS) face significant challenges in path tracking and control during berthing, especially in severe maritime environments.
• Traditional control methods often struggle with the complex dynamics and environmental variations encountered during autonomous operations.

Purpose of the Study:

• To develop an intelligent and accurate berthing system for USS using nonlinear model predictive control (NMPC).
• To enhance the precision of ship berthing control by dynamically adapting to the ship's state and environmental conditions.

Main Methods:

• A 4-DOF (four-degree-of-freedom) ship berthing path tracking method based on NMPC was proposed.
• The Fossen dynamic model (4-DOF: sway, surge, yaw, roll) was established, incorporating moving horizon estimation (MHE).
• A real-time NMPC system was designed to predict trajectories and optimize control inputs.

Main Results:

• Simulations in the Port of Hamburg demonstrated the algorithm's effectiveness.
• Track error was maintained below 2 meters.
• Berthing position error was reduced to only 0.3 meters.

Conclusions:

• The proposed NMPC-based approach offers superior efficiency, generalization, and suitability for autonomous ship berthing.
• It significantly improves attitude control compared to traditional PID and standard NMPC methods, especially in rough seas.
• Provides a theoretical basis and practical recommendations for enhancing the reliability of autonomous ship berthing.