Research On Parameter Identification Of Unmanned Surface Vehicle Maneuverability Based On RLS Method

In recent years,with the exploration of marine resources deeply,unmanned surface vehicle（USV）is favored for its high intelligence,lightweight structure and flexible movement,while more and more researchers have paid attention to the development of marine equipment.The intelligence foundation of USV is to predict its maneuverability with high precision,however,the accuracy of USV maneuverability prediction depends on the accuracy of the established motion model.Therefore,researching the motion model of USV is critical to the intelligent development of USV.Firstly,the response model of USV maneuvering motion is studied in this dissertation,including the first-order nonlinear response model and the second-order nonlinear response model.A parameter identification model based on the recursive least square（RLS）method is designed for response model,and 20° zigzag maneuvering motion simulation tests is carried out to collect the data required for identification,on which the model parameters can be identified.When analyzing the identification results,it is found that some of the parameters of the second-order nonlinear response model have low identification accuracy.A method that optimizing the parameter identification results based on particle swarm optimization（PSO）is further proposed to improve the accuracy of parameter identification of the secondorder nonlinear response model.Maneuvering motion simulation tests are carried out to verify the accuracy of the parameter identification of the first-order nonlinear response model and the effectiveness of the optimization of the parameter identification results of the second-order nonlinear response model.Secondly,the three-degree-of-freedom motion model of the horizontal plane of the USV is studied,including the linear and non-linear models.The RLS algorithm identification model is designed based on the acceleration difference method for the both linear and non-linear motion model respectively,and 20° zigzag maneuvering motion simulation tests is carried out to collect the data required for identification.The inertial hydrodynamic derivative which can’t be identified in the model is estimated by using the slender body theory.Then the identification result of the three-degree-offreedom linear and nonlinear motion model of the USV can be obtained.A parameter drift phenomenon occurred in the parameter identification results of the non-linear models.A differential method based on force and moment was further proposed to redesign the identification model.Based on this model and simulation data,improved parameter identification results were obtained.The identification results before and after the improvement are compared to prove the validity of the method,and a series of maneuvering motion simulation tests are performed to verify the accuracy of the parameter identification results of the linear motion model and the improvement of the parameter identification results of the nonlinear motion model.Finally,the four-degrees-of-freedom of surging、swaying、yawing and rolling motion model of the USV is studied.Considering that the hydrodynamic derivatives in the model have different contributions to maneuverability,the model is simplified without affecting maneuverability.The identification model was designed based on the force and moment difference method,and the experimental data of 20° zigzag maneuvering motion simulation was collected.The identification results were obtained by the RLS identification algorithm.After analyzing the identification results,it was found that some parameter accuracy still had parameter drift.The identification model is redesigned based on acceleration difference method with value-added,and the amplitude of the value-added is decided by the identification result of the first parameter.Based on the redesigned identification model and simulation data identification,the improved results of parameters identification are obtained.The identification results are compared between before and after the improvement to prove the effectiveness of the method.Based on the identification results,a series of maneuvering motion simulation tests are carried out.The simulation results show that the improved identification results can more accurately predict various maneuvering motions of different angles of the USV with a strong generalization ability.