| In today’s era,the level of science and technology is developing rapidly,sensor technology and artificial intelligence technology are widely used,and Unmanned Surface Vehicle(USV)is becoming more and more intelligent and autonomous.The surface USV as a kind of overwater unmanned platform,when performing tasks,due to the influence of load,speed,and navigation environment interference,the maneuverability of USV will change,so it needs an adaptive and robust control model and sophisticated control system as the support of USV motion control.The twin-screw USV boat has good maneuverability and maneuverability.Therefore,in this paper,the iNav-V type twin-screw USV is developed,and the motion control algorithm is researched based on the USV.In order to solve the problem of accurate modeling of small USV,to obtain a mathematical model of ship motion with high accuracy and simple form,and to address the shortcomings of external disturbances and internal uncertainties in ship dynamics modeling,this paper addresses the characteristic model is applied in the field of navigation.The characteristic model of the twin-screw USV is established,and the sensor parameter data is combined to realize the online calculation of the model parameters.In view of the complexity of the water surface environment and the unpredictability of the effect of the tracking of USV when the motor is rotating,the traditional control methods cannot well adapt to external environmental factors and accurately predict the self-control effect of USV.A data-driven USV path tracking control method based on Actor-critic algorithm,which utilizes the behavior-reward scoring mechanism of reinforcement learning to allow USV to pre-learn a set of behavior rules in a simulated environment for use in USV.Make the best control behavior in different environments,and provide a starting point model for real ship training.Designed and implemented the intelligent and automated iNav-V twin-screw USV experimental platform,the boat-based control center was implemented based on the embedded controller,and the shore-based ground station software was developed based on QT.The transmission test verified the stability of the control system and shore-based ground station software,and proved that the experimental platform can meet the requirements of motion control research.In general,the experiment uses the method of migration learning.First,the designed USV path tracking controller model is initially trained in the simulation environment,and then the model is further trained in the real ship.Finally,a model that can adapt to the environment,Can adapt to the control model of the control system.Finally,through the analysis of the results of real ship experiments,it is verified that the USV path tracking controller designed in the actual navigation environment has good control performance and can realize the planned path tracking of the iNav-V USV. |
