The Nonlinear Adaptive Control For Underactuated Surface Vessels

This paper mainly investigates the issues of nonlinear and stable adaptive neural network based control for underactuated surface vessels. The underactuated ship system is a typical second-order nonholonomic dynamic systems. The control problem can't be solved by smooth time invariant feedback controlling. Some nonlinear control methods, developing with the nonholonomic systems problem, can't be used directly for the problem. In practice, the parameters of ship model are usually uncertain and the disturbances of wind, wave and current must be deal with. So the simple onefold control method can't solve such a complicated control problem, It has been paid more attention for searching new methods to fulfill the ship engineering needs in recent years. In order to solve the problem of controlling underactuated surface vessel with model uncertain and environment disturbing, the following research have been completed in this dissertation.1. The stabilizing controller based on exact feedback linearization for ship straight-path tracing, dynamic positioning, and docking have been designed. In order to remove the effect of swaying velocity, the author adopts such a underactuated surface ship motion model designing directly to the virtual input into the design of the kinematics loop, making the system into a first-order nonholonomic system, to overcome the shortcomings of needing exact model parameters from dynamics loop. At the same time, the swaying velocity is the zero dynamics subsystem with global asymptotic stability. The feedback linearization technology is used to design feedback controllers for realizing error signal to achieve closed-loop system asymptotically stable at equilibrium point.2. Aiming at the underactuated ship nonlinear model with uncertainties and outside disturbances, a robust adaptive tracking controller is designed. When external disturbances and model uncertainty are not considered, the controller, designed by applying adaptive technology and cascade system stability analysis theory, makes the closed-loop system error signal globally asymptotical stable at equilibrium. When considering the external disturbances and model are uncertain, the error signal of closed-loop system in the equilibrium point is uniformly ultimately bounded. 3. The controller is designed in condition of unknowing dynamic loop model parameter and relationship of underactuated ship states by means of introducing neural network stable adaptive control. Stable adaptive neural network controller of path following for underactuated ships is designed combining parameter path technique.4. Dynamic positioning and docking control is designed based on path following control. The controller which combines stabilization control and path following control, may lead the underactuated ships following different path to achieve the control objectives. This method can be used to control the ship sailing in restricted water areas.