Research On Parameter Identification Of Maneuvering Motion Model Of USV

With the deepening of the National Marine Power Strategy,in the field of ship,the research demand for unmanned surface vessel(USV)is increasing day by day.With the improvement of marine engine,ship shape optimization design and ship propulsion technology,the design speed of USV is also faster,so its maneuverability requirements are also improved.In order to better study the motion control characteristics of USV and predict its maneuverability accurately,it is very important to establish a more accurate maneuvering motion model.The existing mature methods for obtaining parameters of USV maneuvering motion model are usually time-consuming,laborious,with many restrictions and poor economy.At present,with the rapid development of computer technology,a new system identification method for obtaining model parameters of dynamic system has attracted wide attention,and it has been widely used in the parameter identification of USV maneuvering motion model.In order to overcome the shortcomings of traditional methods and get the parameters of USV maneuvering motion model efficiently and conveniently.Based on the extended Kalman filter and least square method,this paper studied the parameter identification of USV maneuvering motion model,and obtained the parameters of USV maneuvering response model and Abkowitz model.Firstly,the coordinate system of USV motion is introduced.According to the types of USV maneuvering motion models,the USV maneuvering response model and Abkowitz model are introduced respectively.Then,the USV maneuvering response model is discretized by the preceding difference method and transformed into an augmented state vector based on EKF.After that,the parameters of the response model are identified by using the simulation data,and the generalization ability of the identified response model is verified by comparing the simulation results.Then,on the basis of the identification results by the traditional least square(LS)method,a stepwise identification method based on LS is proposed.Based on the correlation analysis of the input data of the identification model,and the discretized response model is simplified to obtain a simplified identification model.And based on the simplified identification model,the identification results of some parameters are independent of each other.Based on this,error correction is carried out step by step.Finally,the parameters of the USV maneuvering response model are identified step by step.By comparing the simulation results of a series of different types of maneuvering motion,it is verified that the stepwise identification method based on LS can better reduce the parameter drift and identify a more accurate response model,which has good generalization ability.Finally,the Mariner ship model is taken as the research object,and the corresponding identification model is constructed based on the Abkowitz model.Combining with the characteristics of the identification model,the identification model is improved by adding an appropriate random variables to reduce the parameter drift.Then the Abkowitz model is identified by using mixed maneuvering motion data,and a series of different types of maneuvering are carried out.The comparison of motion simulation results show that the identification results have good generalization ability.