RBF Neural Network To Optimize ADRC In Ship Track Application In Control

In order to solve the problem of on-line parameter tuning of nonlinear auto disturbance rejection controller and improve the safety and economy of the ship sailing at sea,this paper adopts the auto disturbance rejection control method optimized by radial basis function(RBF)neural network to solve the problem of ship's track control.Auto disturbance rejection control technology is a modern control technology developed after the improvement of PID control technology.The development of artificial intelligence algorithm provides a new idea for auto disturbance rejection control algorithm.The simulation results show that the controller designed in this paper improves the ship's track control effect.In this paper,the MMG model,which is a separate model of ship motion,is used.The hydrodynamic forces and interference forces in the model are analyzed in detail one by one,and the simulation results show that the model error is within the allowable range.For the track control problem of underactuated ship without transverse power plant,the desired bow angle equation is constructed by using hyperbolic tangent function,and the track control problem is transformed into the course keeping control problem.The parameters of conventional NLADRC are difficult to be adjusted and the anti-jamming ability is poor.Inspired by the neural network to adjust the PID controller parameters,this paper combines the RBF neural network with ADRC,designs the ADRC of ship track optimized by RBF neural network,and takes the system input and output as the input of neural network,so that the network output approaches the system output.Two parameters of NLADRC controller,which are equivalent to proportional gain and differential gain in PID control algorithm,are adjusted on line.Finally,taking Shan Hai container ship as the simulation object,using MATLAB to carry out the straight line and curve track keeping control experiments under the sea condition with wind current interference,the simulation results show that the controller has fast tracking speed and high accuracy,and has strong robustness to external interference.