Path Following Control Research Of USV With Time-varying Inertial Array

With the increasing importance of maritime safety and the continuous progress of marine science and technology,unmanned vessel,as an intelligent marine equipment,its study has attracted the attention of all walks of life.Since the marine environment is complex,when the unmanned vessel works,the sailing route needs to be planned in advance,so that the unmanned vessel can move along the route.Equipment on the unmanned vessel varies from one task to another that it performs,this will affect the quality of the unmanned vessel directly.What's more,its additional mass,i.e,the inertia hydrodynamic coefficient,will change during high-speed movement.Besides,because of complex sea condition,the system inertia matrix will become uncertain.Therefore,in consideration of the uncertainty of the system inertial matrix,also in order to eliminating the influence of the interference in marine environment,a backstepping self-adaptive path tracking controller and a neural network based unmanned submarine path tracking controller are designed to improve the tracking performance of the system in this paper.The contents of this paper are as follows:First of all,the North East coordinate system and the hull coordinate system are used to analyze the motion model of the vessel,and the coordinate transformation between the two coordinate systems is calculated.On this basis,the mathematical model of USV is established by analyzing its kinematic and dynamic characteristics.Then,the effects of environmental disturbances were studied and related mathematical models were established.Also,after analyzing changes in the inertial matrix of the unmanned vessel during high-speed moving targetedly,the model simulation experiment was conducted to verify the correctness of the above mathematical model.Secondly,for path tracking of underactuated unmanned vessel,in order to solve the problem that the line-of-sight method only applies to straight-line path tracking,Serret-Frenet(SF)coordinates are used to redesign it,and the line-of-sight method is improved,based on the differential homeomorphism transformation,the tracking error in the North East coordinate system is converted to the error in the SF coordinate system.The guiding controller is designed to transform the path-tracking problem into forward tracking and speed control,thus the relevant control model is simplified.By combining the backstepping method with the adaptive control method,and estimating the inertial matrix and the steady-state interference,the path-tracking controller is designed to eliminate the influence of environmental disturbances and solve the problem about uncertainty of inertial matrix.In the integral guidance algorithm,the drift angle is processed by integral compensation,so that the path tracking task can be completed by the unmanned vessel without navigation system or with a navigation system in a failure state.After simulation research,the purpose of path tracking is achieved effectively with this control algorithm.Thirdly,a backstepping adaptive control method based on robust control is designed considering the variation of model parameters.For the unmanned vessel whose time-varying inertia matrix has nominal value,the model of the unmanned vessel is simplified at first,by processing the model,the changed part of the model is successfully converted to the treatment of time-varying interference,thus the design of the controller is simplified.At the same time,the backstepping adaptive robust controller is designed by assuming the upper bound of interference,so that the improved backstepping adaptive method can be applied to deal with time-varying interference,the path tracking is realized accurately and response time is shortened.Finally,considering that the nominal value of parameters in the model is unknown,and,the change of parameters and the hydrodynamic coefficient when the USV moves with a high speed will affect the system,an RBF neural network adaptive controller is designed,whose middle layer function is the Gauss function.Since the neural network has the characteristic of approximating the indefinite function,it behaves well on solving this problem.A neural network stable adaptive controller is designed for the path tracking of the unmanned airship,in which the Gauss function is selected as the middle layer function,so that the path tracking control of straight line and any curve can be realized under the condition that the parameters of the model is unknown and the external interference exists.