Ship RBF Neural Network Adaptive Sliding Mode Track Control

At present,with the rapid development of artificial intelligence and advanced control technology,the surface ships are becoming more and more intelligent and automated.Surface ships motion control is an important part in the field of ship equipment,which has attached much attention from scholars.The ships are greatly affected by the external environment during sailing,if the external factors are fully considered when designing the ship motion controller,the robust performance of the ship control system can be improved,and it has obvious significance for the ship control system.In the path following process of underactuated ships,there are problems such as internal model uncertainty and external interference.In order to solve these problems,this thesis proposes a sliding mode control algorithm based on radial basis function(RBF)neural network.In the design process of the controller,the reference heading angle is first designed by combining the lateral displacement deviation with the backstepping method.Next,the hyperbolic tangent function is introduced to design the nonlinear sliding surface and the sliding mode control law is obtained by calculation.Finally,RBF neural network is introduced to approximate rudder angle gain,internal model uncertainty and external disturbance uncertainty,as well as the disturbance caused by flow disturbance and transverse velocity,and the stability is analyzed by constructing Lyapunov function.The innovations of this thesis mainly include the following two aspects:(1)In the existing research on ship motion control,the rudder angle gain of the ship is regarded as a fixed value.However,in the actual operation of the ship,the rudder angle gain changes and cannot simply be regarded as a fixed value.In this thesis,the rudder angle gain is regarded as a variable,and the rudder angle gain and the uncertainty items of the internal model and external disturbance are approximated and compensated by the RBF neural network.(2)Considering that the ship is easily affected by the external environmental disturbance and produces lateral displacement,and considering that the RBF neural network has the characteristics of being able to approximate nonlinear functions.In this thesis,the RBF neural network is used to approximate the disturbance caused by flow disturbance and lateral velocity,as well as the uncertainty of internal model and external disturbance.Finally,in order to verify the effectiveness and robustness of the controller designed in this thesis,a ship separation model(Mathematical model group,MMG)is used for simulation in the MATLAB platform.The simulation results show that the controller designed in this thesis can follow the reference straight path and the reference curve path well,and the RBF neural network can also make a good approximation to the rudder angle gain,the uncertainty of the internal model and the uncertainty of the external disturbance,as well as the disturbance caused by the flow disturbance and lateral velocity.In summary,the ship path following controller designed in this thesis has good robustness.