Research On Underwater Acoustic Navigation Filtering Algorithm Under Combined State

Underwater vehicles have been widely used in the fields of marine resource development,deep sea exploration,marine investigation,intelligence reconnaissance and torpedo identification.The normal and efficient work of underwater vehicles is inseparable from high-precision navigation.Underwater acoustic navigation technology can be used for underwater navigation alone for the high precision.It can also provide error calibration for INS and DVL,and effectively control the time cumulative error.Therefore,acoustic navigation technology has become a research hotspot of underwater navigation technology.However,the complexity of the underwater environment and the variability of the carrier’s motion state make underwater acoustic navigation process complicated.In this paper,the high-precision filtering algorithm of underwater acoustic navigation combined linear and nonlinear motion is systematically studied.The main research work and achievements are as follows:1.The research status of underwater acoustic positioning and navigation technology,ocean sound velocity field construction and filtering algorithm is summarized.Three kinds of underwater acoustic navigation principles and common integrated navigation systems are described.The possible errors during acoustic navigation are analyzed from two aspects of system error and accidental error.Finally,several common sound velocity correction methods are described for the sound ray bending.2.In order to construct the ocean sound velocity field,it is necessary to convert the CTD data into SVP data.In order to obtain a high-precision SVP data,this paper compares and analyzes 12 kinds of sound velocity empirical models,and obtains four sound velocity empirical models that can be well fitted in the depth range of 0-200 m,200-1000 m,and1000-2500 m.Then,the sound velocity empirical model is used to convert CTD data into SVP data,then the traditional sound velocity field construction method can be used to obtain the sound velocity field model.3.For the target’s linear or nonlinear motion state,apply the linear(KF)and nonlinear filtering algorithm(EKF and UKF)to estimate the state.The results show that KF can estimate the target’s linear motion.When estimating the target’s nonlinear motion,no matter from which aspect of accuracy,stability and time-consuming,UKF is superior to EKF.4.Aiming at the problems of linear or nonlinear combined state model transformation,kinematics model selection error and abnormal error,the filter selection factor is designed.The experimental results show that the filter selection factor can effectively distinguish linear and nonlinear states.The robust adaptive KF/UKF algorithm for underwater navigation based on innovation vector is designed.The simulation data test results show that the robust adaptive KF/UKF algorithm based on innovation vector can select the filtering algorithm and kinematics model well,and has higher robustness and accuracy.The navigation accuracy of three measured lines with a distance of more than 20 km is better than 50 m.