| Due to the complex physical characteristics of the underwater space environment and the multi-source and dynamic changes of noise,the traditional centralized or sonar array target tracking method cannot achieve high precision underwater positioning and tracking.Due to the characteristics of large monitoring range and flexible deployment,underwater distributed network provides more real-time and effective data support for underwater target state estimation,and makes it possible to track underwater motion trajectory with high precision.In order to reduce the impact of underwater environmental noise on the accuracy of observation information and provide high-precision underwater positioning and tracking services,this paper comprehensively discusses the existing underwater tracking methods,and adopts the simulation method to compare the different methods,and analyzes the advantages and disadvantages of the different methods.In addition,a target tracking algorithm based on particle filter is proposed to improve the accuracy of underwater target tracking.This paper includes the following two aspects:A target tracking algorithm based on distributed perception network is proposed to solve the problem of underwater environmental noise.Considering the underwater environment low signal-to-noise ratio,the target signal and the signal transmission distance is limited adverse factors,such as to reduce the underwater environment noise impact on the observation of the accuracy of the information,this paper proposes a dynamic network resources allocation mechanism,make full use of the network at the level of perception of information,through the deployment of resource nodes enhance the overall network detection ability,To achieve the accurate acquisition of target observation information in the underwater monitoring network.In addition,a particle filter algorithm based on distributed state fusion is designed to improve the influence of various observation noises on the tracking accuracy of a single target without considering the data association and information false positives in multi-target tracking.Simulation results show that,compared with the traditional algorithm,the proposed method can achieve better tracking accuracy and higher stability in a variety of non-Gaussian noise environments and in the case of nonlinear target motion.Aiming at the intersection problem of multiple targets in a small range,a two-layer particle filter algorithm is introduced in this paper to increase the tracking accuracy of underwater multiple targets.In this paper,the underwater multi-target tracking problem in a small area is studied.In addition to the influence of non-Gaussian observation noise and low signal-to-noise ratio,the problem of false alarm interference caused by mismatching of observation information and similar features is also studied.Especially when multiple targets have cross movements,the above problems will become very serious,which may eventually lead to the explosion of calculation amount of the algorithm and deviation of tracking trajectory.In this paper,a two-layer particle filter algorithm based on distributed probability fusion is proposed.In the problem of multi-target trajectory intersection,the algorithm can obtain accurate data association according to the target's historical track and the observation particle's state information,and then obtain more effective information.In the fusion layer,an independent particle filter is applied to the preprocessed spatial observation information to improve the accuracy of the observation information.The simulation results show that the proposed algorithm can provide higher positioning accuracy and state estimation information when multiple targets are near each other in a small area. |
PDF Full Text Request