Research On Multi-UUVs Cooperative Maneuvering Target Tracking Method Based On Bearing-only Detection

Unmanned Underwater Vehicle(UUV)to achieve stable tracking of targets is often the basis for other tasks.Affected by the UUV’s detection range and maneuvering characteristics,a single UUV is prone to target escape during target tracking tasks.While making up for the above shortcomings,the multi-UUV system has more advantages in target state estimation accuracy and system robustness.In this paper,the research on multi-UUV coordinated maneuvering target tracking based on pure azimuth detection is carried out,and the following researches are mainly carried out:Firstly,it analyzes the multi-UUV cooperative maneuvering target tracking,and briefly summarizes the characteristics of the multi-UUV cooperative maneuvering target tracking.And according to the different tracking methods,the multi-UUV coordinated maneuvering target tracking mode is divided into open-loop tracking and closed-loop tracking.Then the UUV dynamics model,target motion model and passive sonar-only azimuth detection model are established.Secondly,the research is carried out on the state estimation method of maneuvering target based on pure azimuth detection.The simulation compares the performance of extended Kalman filter(EKF),unscented Kalman filter(UKF),and particle filter(PF)under pure azimuth detection.Since the motion model of a maneuvering target is not a single model,it is usually a combination of several models.Therefore,the estimation algorithm of a single model often cannot accurately estimate the maneuvering target.Therefore,this paper studies the EKF method based on interactive multimodality(IMM).Since the filtering algorithm is very sensitive to the initial filtering value,it is particularly important to set an accurate initial filtering value.However,in actual application scenarios,the initial value of the target cannot be known in advance.In response to the above problems,this paper proposes a method based on the least squares method.The target initial state estimation method is used to determine the initial filter value of the state estimation algorithm.Thirdly,research has been carried out on the fusion state estimation problem of multiUUV coordinated maneuvering target tracking.Compared with single UUV target tracking,multi-UUV target tracking increases the dimension of measurement information.In order to better utilize the measurement information of multiple UUVs,Several fusion structures were compared.According to the characteristics of multiple UUV systems and the advantages and disadvantages of different fusion structures,a distributed fusion structure was selected.And using simulation verification to compare three distributed fusion algorithms,and finally choose the best performance information matrix fusion method as the fusion method of this article.Finally,research is carried out on the optimization method of multi-UUV cooperative maneuvering target tracking path,in order to realize the optimal maneuvering of multiple UUVs on the basis of safe tracking and maintaining tracking to reduce the target state estimation error.In this paper,the path optimization performance function is established based on the Fisher matrix(FIM),and the rolling solution framework is established based on the particle swarm optimization method.In addition,considering the impact of the large delay of underwater sound,the frequency of UUV path optimization cannot be higher than the communication frequency,so the solution time domain is divided,and the operations that should be performed in each time period and the content of information interaction are clarified.