Research On Cooperative Positioning Algorithm For Autonomous Underwater Vehicles

With the deep development of marine technology,the role of autonomous underwater vehicle(AUV)coordination system in military applications is becoming more and more important.The high-precision positioning of autonomous underwater vehicles is one of the necessary conditions for the successful execution of assignment.However,the underwater environment is not conducive to the propagation of electromagnetic wave signals,which makes it difficult for satellite navigation systems to accurately locate autonomous underwater vehicles.Due to the high price,the high-precision inertial navigation system is not conducive to large-scale applications.Part of AUVs equipped with high-precision inertial navigation system is used as the leader AUVs.The information is transmitted and the relative distance is measured by underwater acoustic equipment.Then the data fusion technology is used to improve the positioning ability of the other following AUVs equipped with low precision navigation system.Therefore,it is of great significance to carry out the Research on cooperative positioning technology of autonomous underwater vehicles based on underwater acoustic ranging.This topic focuses on the high-precision positioning requirements of autonomous underwater vehicles,and develops a leader-follower autonomous underwater vehicle cooperative positioning technology based on underwater acoustic ranging.This paper deeply analyzes the observability problem of the cooperative positioning system,and proposes a cooperative positioning algorithm based on the factor graph and the sum product algorithm in Gaussian noise environment.For the heavy tail noise problem of underwater acoustic ranging,a cooperative positioning algorithm based on the factor graph and the maximum correntropy is designed,at the same time,the error parameter identification algorithm of follower AUV is proposed for the first time to improve high-precision positioning capability of multiple autonomous underwater vehicles.The main contents of this dissertation are summarized as follows:Firstly,for the autonomous underwater vehicles cooperative positioning system,a mathematical model of the cooperative positioning system is established and the extended Kalman filter is used to realize the cooperative positioning technology,and the cooperative positioning technology is verified through simulation experiments.For the observability of the autonomous underwater vehicle cooperative positioning system.A cooperative coordinate system to decouple the relative motion of the leader AUV and the follower AUV is proposed,In the case of considering the control input in the cooperative coordinate system,the article separately studied the observability of the single leader AUV cooperative positioning system and the two leader AUV cooperative positioning systems,and defined the observable degree of the system.Based on the above analysis,a simulation experiment is designed to verify the analysis results of observability.The simulation results show that when the observable degree of the system increases,the positioning error will converge to a smaller range.The simulation results prove the correctness of the analysis results.Secondly,for the problem of large positioning error and complicated computation of traditional algorithms of the cooperative positioning system in Gaussian Noise environment,the factor graph model of the cooperative positioning system is established,and a cooperative positioning algorithm based on the factor graph and the sum product algorithm is proposed.The expectation and variance are passed between the nodes in the factor graph to complete the follower AUV positioning error correction.Then,the conversion matrix is introduced into the factor graph model.An improved cooperative positioning algorithm is proposed to effectively avoid the positioning error caused by the same coordinates of the follower AUV and leader AUV,and reduce the computational complexity of the algorithm.The proposed algorithm is verified by the simulation experiment and the offline data of the actual ship experiment,and the result proves the effectiveness of the proposed cooperative positioning algorithm.Thirdly,for the problem that the positioning accuracy is severely affected by HeavyTailed Noise in the observation of cooperative positioning system,The reason of HeavyTailed distribution of the observation error in underwater acoustic ranging is analyzed,and a cooperative positioning algorithm based on factor graph and maximum correntropy is proposed.And the maximum correntropy criterion is introduced into the factor graph model as a cost function,and an adaptive kernel width algorithm based on observation error is designed,which effectively suppresses the influence of heavy tail noise on the positioning accuracy of the algorithm.On this basis,a cooperative positioning algorithm that uses sliding window to construct observation information is proposed,and a kernel width determination method based on median filtering is designed.An adaptive window algorithm for adjusting the window size based on the proportion of the outliers in the window is established.To verify the effectiveness of the algorithm,the two algorithms are verified using simulation experiments and offline data of real ship experiments.The experimental results show that both algorithms can effectively reduce the positioning error when the observation contains heavy tail noise.Finally,for the problem that follower AUVs have speed error and course error in cooperative positioning system with relative distance as observation,the influence of the two process errors on the positioning accuracy of follower AUVs are analyzed in detail for the first time,and a factor graph model of error parameter identification algorithm is proposed.Then,error parameter identification algorithms based on Gaussian noise and heavy tailed noise are proposed respectively.Under the condition of Gaussian noise,the expectation and variance are transferred among the nodes of the factor graph to complete the parameter identification of velocity error and heading error.Under the condition of heavy tailed noise,the correntropy is used as the cost function to update the information among the nodes of the factor graph,it suppresses the influence of process error on positioning accuracy.In order to verify the effectiveness of the algorithm,the off-line data of simulation experiment and real ship experiment are used to verify the two error parameter identification algorithms.The results show that the two algorithms can effectively reduce the positioning error of follower AUV,especially in the autonomous positioning of follower AUV.