Simulation Research On Ship Auto-Berthing Based On Active Disturbance Rejection-Neural Network Control

Ship berthing is the most important part of ship navigation.The captain must have rich experience to drive the ship to complete the berthing task of the ship.Due to the proposal and development of intelligent ships,the auto-berthing of ships has also become an urgent issue in the field of ship navigation.Based on reading literatures,this paper proposes a method combining active disturbance rejection control and neural network to complete the berthing task of the ship.As the ship sails in the port at low speed and shallow water,and the ship's force changes greatly compared to normal speed and deep water,the ship controller cannot accurately complete the navigation control tasks in port.Based on the characteristics of the ship's low-speed and shallow-water conditions in port,the following tasks have been completed:(1)Establish the MMG motion model of the ship by querying the data,and modify the ship model in low speed and shallow water conditions,respectively.Taking into account the external disturbances to the ship's navigation in port,the effects of wind,wave,and current on the ship's navigation are queried.The wind,wave,and current models are established respectively.(2)Ship's in-port navigation has the characteristics of inaccurate models,which makes it impossible to design accurate control algorithms,and' the ship is relatively affected by environmental interference in low-speed shallow water.And the active disturbance rejection control algorithm has the characteristics of low accuracy of the control object,excellent anti-interference performance and simple design.In this paper,the ship model is simplified to a third-order Nomoto model,and the change of the draught ratio and speed of the ship is related to the ship's maneuvering index.The design of the active disturbance rejection control algorithm is used to complete the course keeping control of the ship navigation in port.(3)Analyze the control characteristics of under-actuated ships,and divide the ship's track control into two parts:ship's course control and position control.By designing active disturbance rejection controllers,the ship's straight track control is completed.(4)By collecting the berthing information of the ship in different initial states,the neural network is trained to complete the control of the propeller speed during the berthing of the ship to control the ship's Speed,and the effectiveness of the control is verified by simulation.In this paper,the active disturbance rejection controller is used to complete the ship's track control during berthing.By collecting the berthing information of the ship in different initial states,the neural network is trained to complete the control of the propeller speed during the berthing of the ship to control the ship's Speed,and the effectiveness of the control is verified by simulation.This paper verifies the effectiveness of the active disturbance rejection control algorithm in the navigation control of a ship's navigation in port,and uses a combination of active disturbance rejection control and neural network control to complete the berthing control task of the ship.