Virtual Guidance-Based Coordinated Tracking Control of Multi-Autonomous Underwater Vehicles Using Composite Neural...

Area of Science:

• Robotics and Control Systems
• Marine Engineering
• Artificial Intelligence

Background:

• Coordinated control of multiple autonomous underwater vehicles (multi-AUVs) is challenging due to nonlinear dynamics and system uncertainties.
• Existing learning methods often prioritize stability over performance metrics.

Purpose of the Study:

• To develop a virtual leader-based coordinated controller for nonlinear multi-AUVs.
• To improve tracking performance and uncertainty approximation accuracy in multi-AUV systems.

Main Methods:

• A virtual leader approach is employed, with control commands based on relative positions to the virtual leader.
• A back-stepping control scheme is combined with an online data-based learning approach for uncertainty approximation.
• A novel learning performance index, utilizing online data, is integrated into the neural weight update law.

Main Results:

• The proposed controller demonstrates enhanced tracking performance compared to previous methods.
• Improved approximation accuracy for system uncertainties was achieved.
• Closed-loop system stability was validated using the Lyapunov approach.

Conclusions:

• The virtual leader-based coordinated controller effectively manages nonlinear multi-AUVs with uncertainties.
• The novel learning performance index enhances control system adaptability and precision.
• The method offers a promising solution for advanced multi-AUV coordination and formation tasks.