Multi-Observer Fusion Based Minimal-Sensor Adaptive Control for Ship Dynamic Positioning Systems.

Area of Science:

• Marine Engineering
• Control Systems Engineering
• Robotics

Background:

• Dynamic positioning (DP) systems are crucial for marine vessel station-keeping.
• Traditional DP systems often rely on numerous sensors, increasing cost and complexity.
• Handling model uncertainties and input constraints remains a challenge in DP control.

Purpose of the Study:

• To propose an adaptive dynamic positioning (DP) control method with minimal sensor requirements.
• To enhance the robustness and accuracy of DP systems against disturbances and uncertainties.
• To validate the proposed method's performance through simulations.

Main Methods:

• A multi-observer fusion architecture combining sliding mode and finite-time disturbance observers.
• A single-parameter learning neural network to manage model uncertainties.
• Auxiliary dynamic systems to address input saturation and thruster dynamics.

Main Results:

• The observer-fusion strategy demonstrated stable performance.
• Effective online estimation of system states and disturbances.
• Successful handling of model uncertainties and input saturation constraints.
• Simulation results validated superior performance over traditional methods.

Conclusions:

• The proposed adaptive DP control method offers a robust and sensor-efficient solution.
• The multi-observer fusion architecture effectively addresses complex control challenges.
• This approach enhances the reliability and applicability of dynamic positioning systems.