Research On The Nonlinear Adaptive Filter Of Ship Dynamic Positioning System

Ship dynamic positioning system is refers to the ship which didn’t need the traditionalanchoring system．It could resisted from wind, wave and flow interference to make the shipkeep in anchor point nearby．Compared to other engineering ship, dynamic positioningsystem was a very good solution to solve the traditional deep which hard to locate, andprovide the safeguard to deep working．The movement of the ship at sea basically can be divided by the resultant force into thehigh frequency motion and the low frequency motion．The high frequency motion isprovided by first order wave force which caused the ship reciprocating oscillation, this kindof motion damage the ship’s host is very big, and for the ship motion influence is lesser, thatcould be filtered．Low frequency motion is produced by second order wave forces, whichinfluence to the ship’s motion is bigger, when low frequency motion on the ship’sdisplacement affects more than a set category, ship host will be started produce a forcewhich counteractive to second order wave force that make the ship back to the set upposition．Therefore, we found that filtering plays an important role in dynamic positioningsystem, its practical significance is very obvious．Based on the ton trailing suction hopper dredger as the article research object,established a dredger movement mathematical model, including the high frequency motionmodel, the low frequency motion model, measurement model, the environment disturbingforce model and propeller model and so on, combined with the ship dynamic positioningmodel designed a kind of nonlinear adaptive filter．At present most used filter is kalmanfilter, kalman filter is mainly used to solve linear filtering problem．Direct at nonlinearfiltering, appeared to the extended kalman filter, but the filter is difficult to calculate jacobimatrix, especially for big system is not easy to solve．The design of the center differentialkalman filter is based on linear regression change project transformation, compared to solvenonlinear filtering extended kalman, which can obtain the smaller linearization error, and atthe same time avoid the Jacobi matrix calculation．Through the verification, the design filtercan be attached to the ship’s low frequency motion position from the output informationwhich contain the measurement noise, bring certain reference to nonlinear ship motionfiltering, that contribute to further learn nonlinear filtering issues．Finally, this paper makes the simulation experiment, the first research is trajectory tracking simulation, through the simulation experiments we can show that the design of thefilter can estimate the position of the ship from with wind, wave and current noiseinterference．Then by joining a simple controller, carries on the fixed point positioning, bysetting a initial value to verify whether target ship reaches the specified location in a shorttime, realized fixed point positioning．The article main work focused on the study ofnonlinear adaptive filter, handle the nonlinear model, and based on this, used centerdifferential kalman filtering algorithm, through the simulation experiments, show that thisalgorithm basically conform to the expected．...