Research On Autonomous Motion Control Of Cargo Ships Based On Intelligent Optimization Prediction Algorithms

With the integration and application of high and new technologies such as artificial intelligence,automatic control and data science and the shipping industry,ships are developing in the direction of traditional driving to enhanced driving,assisted driving,remote driving and automatic driving.Fewer people and intelligent of ships have become an inevitable trend.How to enhance the stability of control to ensure the safety of ship intelligent navigation system has become the key scientific problem involved,and how to improve the accuracy of ship control to enhance its adaptability to complex environment has become the focus of academic research.There are many factors that affect the accuracy of ship motion control,such as the parameters of the controller,the model of the research object,the robustness and adaptability of the controller.This thesis introduces the idea of intelligent optimization algorithm and prediction,and studies the heading control,path following,autonomous docking and undocking control of cargo ships.It mainly solves the problems of difficult adjustment of controller parameters,difficult to obtain accurate ship model parameters and poor adaptability of the controller in the motion control of cargo ships.The research work and innovation of this thesis are as follows:(1)Aiming at the problems of difficult parameters adjustment and slow convergence of control process,based on the Beetle Antennae Search(BAS)algorithm,this thesis studies how to enhance its convergence and reduce the probability of falling into local optimization,and puts forward the Antenna Mutation Beetle Swarm(AMBS)algorithm and Antenna Mutation Beetle Swarm-Predictive(AMBS-P)algorithm.They improve the efficiency of controller parameters adjustment and accelerate the convergence speed of control.Based on the analysis of the current research status of ship motion control at home and abroad,the intelligent optimization prediction algorithm is used alone or combined with the other algorithm to improve the adaptability of ship motion control.In order to verify the application effect of the combination of intelligent optimization prediction algorithm and traditional algorithm,AMBS-PID controller is designed.It can reduce the workload of manual parameter adjustment and improve the accuracy of the controller.In order to verify the effect of using the algorithm alone,the prediction idea is introduced,and a ship docking controller based on AMBS-P algorithm is designed.(2)Aiming at the situation that the prediction algorithm needs the ship model and the accurate ship dynamics model is not easy to obtain,a method combining model identification with AMBS-P or Antenna Mutation Beetle Swarm-Predictive-Reinforcement Learning(AMBS-P-RL)algorithm is proposed to realize the accurate motion control of the ship.The accurate model reflecting the real ship dynamics is highly nonlinear and complex,especially when the experimental conditions are limited,it is difficult to obtain the model parameters,which makes the construction of the accurate model difficult.According to the need of controller with prediction idea for high-precision model,a controller based on model identification and AMBS-P algorithm is designed.In order to further enhance its adaptability,the reinforcement learning idea is introduced.In the process of control,the ship actions are learned interactively with the environment to maximize the reward obtained in each step of control,then the rudder order correction value is obtained to compensate the motion control error.(3)In view of the large error when the heading angle changes greatly,the“optimal switching strategy of motion control method” is proposed and verified by experiments,which effectively reduces the ship motion control error.In order to get the appropriate switching strategy of motion control method,this thesis verifies the application effect of each algorithm through experiments.Firstly,the model ship experimental platform is built.Based on the experimental platform,the cargo ship heading and path following control experiments of controllers based on PID,AMBS-PID,AMBS-P,AMBS-P-RL algorithms are carried out respectively.Secondly,based on the analysis of the application characteristics of each algorithm,the controller with stronger convergence is set to switch autonomously when the ship needs to change its heading greatly.Taking the container model ship as the experimental object,the application of three controllers alone or in combination in four reference heading and two reference trajectory scenarios under the conditions of different wind speeds of level 1-5 is obtained.Model ship experiments in various scenarios verify the adaptability of the algorithm.The implementation of switching strategy can effectively reduce the control error.It can simulate the process that the control system independently selects the appropriate algorithm to realize navigation control according to the ship motion state and navigation scene,which is helpful to enhance the cargo ship’s ability to deal with various environments.To sum up,taking three model ships whose prototypes are oil tankers,tugs and container ships as the research object,it is gradually verified from simulation analysis to model ship test that the control algorithm proposed has the advantages of simple parameter adjustment,fast calculation speed and strong adaptability.Combined with model identification,it is helpful to improve the problem of low motion control accuracy caused by the difficulty of adjusting controller parameters and the difficulty of obtaining accurate ship model parameters.The proposal of “optimal switching strategy of motion control method” reduces the error of ship steering control,enhances the adaptability of ship control system and improves the robustness of ship autonomous navigation to various environments.