Research On Nonlinear Observer Design For Dynamic Positioning Ships

Dynamic Positioning Systems(DP systems)is the necessary systems for ships when they work in the complex sea environment.The ships with DP system can maintain its desired position and heading or follow the pre-determined trajectory by the use of thrusters.While the precision for DP systems is easily to be effected by many situations,so when designing DP control systems for ship,the observer is needed to separate its motion into low-frequency(LF)and wave frequency(WF)apart,filter out the wave-frequency(WF)motion and defend the outside disturbance before they get into the feedback loop,then produce LF position and heading estimates of ship.The paper adopts the sliding mode control theory to design observer for DP systems with unknown uncertain terms.The observer for ship DP systems is used to filter out the WF motion and estimate the LF motion.So a mathematical model of ship dynamic positioning system is firstly established,including the ship LF motion model,WF motion model and the dynamic postioning measurement model.Furthermore,for better study the marine environment's effect,its mathematical model including the wind,wave and current are established as well.To be worthly mentioned,the Coriolis and nonlinear damping terms are as the unknown terms to be considered into the mathematic model of ship LF motion.For ship dynamic position system mathematical model with uncertainties contained,the sliding mode control theory is introduced into the observer design of Ship Dynamic Positioning System.Due to the DP system does not meet the required matching conditions of traditional sliding mode observer design.Then the paper uses the method of constructing auxiliary output to reconstruct the system state space form,and design a high-gain observer to estimate the auxiliary output.Then construct slip sliding mode surface and design strategies on this base.Meanwhile,by the anylysis from Lyapunov stability theory,it is known that the stability of error system is related to the sliding mode gain,the gain should not smaller than the upper bound of unknown uncertain term.Under the assumption that the upper bound of unknown uncertain term is known,the sliding mode observer is designed and simulated.Otherwise,much larger gains are selected in the simulation.It proved that although the systems' robustness needs to be guaranteed,too large gains should not be selected,because it may cause the serious chattering estimated state from the observer.Under normal circumstances,the upper bound of unknown uncertain term in ship DP systems is unknown.In order to prevent chattering or bad robustness caused by two big or small gains,the RBF neural network and Lyapunov stability theory are combined to estimate the upper bound of the unknown uncertain term.An adaptive robust sliding mode observer for ships is designed,and the stability of error systems is certified.Observer is designed for DP control systems,so the simulation of adaptive sliding mode observer with controller is taken at last.The results confirm the validity of the sliding mode observer for DP system with uncertainties.