Online recorded data-based finite-time composite neural trajectory tracking control for underactuated MSVs

Area of Science:

• Marine Engineering
• Control Systems
• Robotics

Background:

• Underactuated marine surface vessels (MSVs) face challenges with uncertain dynamics and external disturbances.
• Existing control methods struggle to achieve precise trajectory tracking under these conditions.
• The line-of-sight (LOS) method is often used to address the underactuation problem in MSVs.

Purpose of the Study:

• To develop a finite-time composite neural control scheme for underactuated MSVs.
• To enhance trajectory tracking performance by addressing uncertain dynamics and time-varying external disturbances.
• To improve the learning capability of neural networks (NNs) through online recorded data.

Main Methods:

• Utilized composite neural networks (NNs) to approximate uncertain MSV dynamics.
• Introduced a modified prediction error signal based on online recorded data.
• Designed a disturbance observer to estimate compound disturbances, including NN approximation errors and external factors.
• Employed a smooth function and Lyapunov approach for stability analysis and finite-time control.

Main Results:

• Achieved finite-time composite neural trajectory tracking control for MSVs.
• Demonstrated improved NN learning ability via a weight updating law driven by tracking and prediction errors.
• Verified the effectiveness of the proposed control scheme through simulation tests on an MSV.

Conclusions:

• The proposed online data-based composite neural finite-time control scheme effectively manages uncertain dynamics and external disturbances in MSVs.
• The control system guarantees stability and ensures trajectory tracking errors converge to a small residual set in finite time.
• The method offers a robust solution for enhancing the navigation and control precision of underactuated marine surface vessels.