Cooperative control for a ROV-based deep-sea mining vehicle with learned uncertain nonlinear dynamics

Area of Science:

• Robotics
• Ocean Engineering
• Control Systems

Background:

• Traditional deep-sea mining vehicles face significant challenges with track slippage, limiting operational efficiency and reliability.
• Existing systems often struggle with precise control in complex underwater environments.

Purpose of the Study:

• To propose an enhanced remotely-operated vehicle (ROV)-based deep-sea mining system (RDMV) to overcome track slippage limitations.
• To develop a cooperative control strategy for improved traction and sinking control of the RDMV.
• To investigate a learning-based model predictive control (LMPC) for robust disturbance rejection.

Main Methods:

• Established the dynamic model of the ROV-based Deep-sea Mining Vehicle (RDMV) using Lagrangian mechanics.
• Developed a distributed model predictive control (DMPC) for virtual speed control laws.
• Implemented a learning-based model predictive control (LMPC) with a Kinky Inference (KI) prediction function for optimal control input and disturbance estimation.

Main Results:

• The LMPC controller demonstrated feasibility and superiority in individual ROV motion control.
• Numerical simulations confirmed the effectiveness of the cooperative control strategy for the RDMV.
• The proposed system successfully addresses the track slippage bottleneck in deep-sea mining.

Conclusions:

• The novel ROV-based deep-sea mining system offers a superior alternative to traditional tracked vehicles.
• The cooperative control strategy and LMPC controller provide robust and efficient operation for deep-sea mining robots.
• This advancement enhances the potential for reliable and productive deep-sea resource extraction.