Dynamics Model Multi-sensor Fusion Online Identification Method Research Of AUV Under The Complex Condition

With the continuous improvement of the economic level of our country,the voice of the nation and the people for building a strong nation in the ocean is increasing.The construction of a powerful ocean nation can not be separated from the observation of the marine environment and the exploration and development of marine resources.And underwater vehicle is one of the important means of achieving these goals.When an underwater vehicle moves in a complex underwater environment,its dynamic parameters will change due to changes in the load it receives.The measurement data during the movement of the underwater vehicle is also easily contaminated by various noises such as Gaussian white noise,complex white noise(white noises in the coefficient matrix and observation vector),colored noise,etc.,the accuracy of the underwater vehicle's motion and control system may also be affected by the complex environment.In order to make real-time corrections to the underwater vehicle 's motion and control system to make it more adaptable to the underwater complex environment,it is necessary to obtain a high-precision dynamic model that can continuously update with the changes of the underwater vehicle's dynamic parameters.The research content of this dissertation is aimed at the parameter identification of underwater vehicle dynamics model under complex environment.In this paper,to meet the real-time and high precision requirements in the online identification problem,the accuracy of parameter identification of underwater vehicle dynamic model is improved rapidly by fusion of measurement data of multiple sensors.The main research contents of this article are as follows:1)Firstly,the kinematics model of the underwater vehicle was established and introduced.Then the dynamic model of the underwater vehicle was introduced.The relevant characteristics of the underwater vehicle were explained and the dynamic model was simplified.2)The content and identification steps of the parameter identification of the underwater vehicle dynamics model are introduced.The identification principles of some identification algorithm are described.Then these algorithms are applied to the identification of the underwater vehicle dynamics model,and the identification results of these algorithms are compared in different noise environments.3)The parameter online identification problem is studied under the conditions that both the coefficient matrix and the observation vector of the underwater vehicle dynamic model contain white noise.The estimation results of the fusion of least squares algorithm(RLS_F)in complex white noise environment and its theoretical deviation with real value is obtained through theoretical deduction.A recursive total least squares online identification algorithm based on multi-sensor fusion(RTLS_F)is proposed.The research content was verified by simulation experiments.4)The parameter online identification problem under the condition that the observation vector of underwater vehicle dynamic model contains colored noise is studied.The principle of augmented least squares algorithm is introduced,and the multi-sensor fusion of augmented least squares algorithm(RELS_F)is proposed through theoretical derivation in order to quickly improve the accuracy of parameter online identification,the performance of RELS_F algorithm is verified by simulation experiments.