Research On Multiple Model Adaptive Observer In Dynamic Positioning

With the fully development of land and Offshore resources, It is getting more and more attention to the development and utilization of deep-sea resources. Since the complication of deep-sea environment, for ship track-keeping and fixed-point operation, the installation of Dynamic Positioning Systems is necessary. DP system can only rely on its own propulsion to eliminate the interference of the external environment, in order to precisely control the position of the ship or tracking,more accurate observer is required to eliminate the measurement noise and estimate the information such as ship's position. An adaptive filtering method is described in this paper to solve filtering and state estimation problem of ship dynamic positioning system in slowly changing sea conditions.The observer of ship dynamic positioning system need to filter out high frequency information of ship, estimate the low-frequency location of the ship. Under reasonable assumptions, a mathematical model of ship dynamic positioning system is established,including the mathematical model of ship high frequency and low frequency linear superposition model and dynamic positioning system measurement model. Furthermore, for better study and simulation of the observer, a marine environment mathematical model including wind, wave and current is established.By comparing a variety of filtering and state estimation method, this article focuses on nonlinear passive observer of the ship. The study of the nonlinear passive' observer is corresponding to ship motion which is a nonlinear, uncertain system. Measurement noise and high frequency information of the ship motion can be efficiently filtered in a fixed sea conditions, hence the speed of ship can be estimated. But the real sea state is slowly changing,wave filtering parameters of observer should also follow sea state changes when the sea condition changed. To solve the problem of observer adaptive filter parameters adjustment, this paper presents a multiple model adaptive observer. The MMAO consists of a set of nonlinear passive observer with different wave filtering parameters and a dynamic weighting signal generator. In changing sea conditions,this observer can automatically switch to sub-observer which match the sea state, filter out measurement noise and high frequency information of ship motion efficiently, estimate the accurate low frequency location of the ship. Finally, by using computer simulation, verify the effectiveness of the multiple model adaptive observer.Considering the multiple model adaptive observer contains multiple sub-observers,each sub-observer has more than one parameters to adjust, the adjustment of parameter are both complicated and tedious, and the performance of the system directly depend on the quality of the parameter adjustment, this paper presents the maximum minimum adaptive ant system optimization to optimize the problem, and verify the validity of MMAAS by using computer simulation.