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

To overcome the bottleneck problem of the track slippage of the tracked mining vehicle in the traditional deep-sea mining system, this paper proposes an enhanced remotely-operated vehicle (ROV)-based deep-sea mining system. A ROV-based Deep-sea Mining Vehicle (RDMV), consisting of two ROVs and a mining robot (MRT), is instead of the traditional tracked Deep-sea mining vehicle. Firstly, the dynamic model of the RDMV as a control object is established based on Lagrangian function. Secondly, a cooperative control strategy is proposed for traction and sinking control of the RDMV. A distributed model predictive control (DMPC)-based controller is developed to obtain virtual speed control laws to meet the control objects. To track the virtual speed control laws, a learning-based model predictive control (LMPC)-based controller is investigated to compute the ROVs' optimal control input, where a Kinky Inference (KI) prediction function is introduced in the state transition model to estimate the unknown external disturbances under random noise. Finally, the feasibility and the superiority of the LMPC controller is preliminarily verified in a degenerate individual motion control of a ROV, and then the cooperative control strategy is proven to be effective through numerical simulations.