Intelligence Ship Handling And Control Of Unmanned Surface Vehicle Under High Sea Conditions

Unmanned Surface Vehicle(USV)is an important maritime unmanned intelligent equipment,which can play an important role in the military and civilian fields,and is a hot spot in the research of major powers in the world today.Navigation,guidance,and control system(GNC)is the key technology of USV,and it is the basis for USV to perform maritime missions autonomously.Traditional GNC system research is usually limited to quiet sea conditions and small sea conditions.Based on simplified dynamic models,line-of-sight(LOS)and other trajectory control methods are used,and sea wind,waves,and currents are not considered.The impact of USV navigation does not consider the encounter angle and maneuvering problems of the USV in large waves,and cannot adapt to navigation tasks under high sea conditions,which limits the application scenarios of the USV.With the rapid development of artificial intelligence technology in recent years,further improving the intelligence level of USV is of great significance to the development of our country’s marine industry.In this regard,this thesis combines traditional control theory with artificial intelligence reinforcement learning methods,and conducts research on USV intelligent ship handling,decision-making,and control issues under high sea conditions,including the following aspects:(1)The dynamic model of USV under high sea conditions is established.In view of the existing USV dynamics model formula dispersion,numerous parameters,application difficulties,etc.,this paper summarizes in detail the horizontal threedegree-of-freedom USV dynamics model suitable for high sea conditions,and distinguishes the heading,heading,drift angle,etc.Basic concepts,detailed descriptions of the calculation methods of sea inertial force,viscous force,sea wind force,first-order wave force,and ocean current force.The detailed calculation process is given by taking the 1.255 m ship model as an example to provide for subsequent research.The basic basis.(2)A USV attitude control algorithm considering wind,wave and current compensation is designed.Aiming at the USV attitude control problem under high sea conditions,the algorithm first compensates for the sea-wind moment through the feedforward control law,thereby transforming the USV attitude control problem into the stabilization problem of the first-order Nomoto equation;then based on the first-order Nomoto equation,through the root The trajectory and Bode diagram are designed with a feedback control law that takes into account the speed change.The simulation results show that adding the feedforward control law can effectively improve the USV attitude control accuracy.(3)An intelligent ship maneuvering decision algorithm based on Deep Qlearning(DQN)is designed.A method of introducing the safety evaluation mechanism into the USV dynamics model is proposed to make up for the defect that the three-degree-of-freedom USV dynamic model of the horizontal plane does not consider the impact of the encounter angle on the roll and pitch of the hull;the intelligence based on the DQN reinforcement learning method is designed Ship handling decision algorithm,detailed introduction of the principle of DQN algorithm in the thesis and the design of action space,state space,reward function and algorithm training process.Finally,the trained intelligent method is simulated and analyzed under high sea conditions.Compared with the traditional LOS lineof-sight guidance method,the intelligent method can autonomously execute the Zshaped ship strategy,and sail safely in the way of alternating starboard and starboard waves.Reach the target point.