Research On Cooperative Target Tracking For Unmanned Underwater Vehicles

Unmanned underwater vehicle(UUV)target tracking is one of the most important independent capacity,ensures the completion of various underwater missions,thus to research it is of great significance.This thesis focuses on the target tracking problems in complex environment,together with the problems of target state estimation and vehicle tracking motion control.The following researches are conducted:First,analyze the solution of target tracking problem,and divide it into two major problems: target state estimation and tracking motion control.On the basis of above researches,mathematical models of the components of target tracking,including tracking platforms,sensors,motion targets,are developed separately.Second,according to the situations with single target and multi-targets,different solutions are designed.As for the state estimation of a single target tracking,on the basis of Bayes estimation theory,this thesis analyzes the Kalman Filter,and the Extended Kalman Filter(EKF),Unscented Kalman Filter(UKF)and Particle Filter(PF)under its nonlinear conditions.For the problem of cooperate tracking with multi-UUVs,disign federated Kalman filter to fuse the state estimation.Meanwhile,the situations of one single target under different filters are emulated in order to compare the differences of calculation complexity,scope of application and errors in state estimation.Simulation experiment is completed for multi-UUVs cooperate target tracking.The result shows that cooperate target tracking works better than tracking with single sensor.As for the state estimation of multi-targets,the major difficulty is the complex condition and changing numbers of targets.On account of that,this thesis analyzes the Probability Hypothesis Density(PHD)filter based on the random finite sets theory,and Cardinalized Probability Hypothesis Density(CPHD)filter on account of the target number,hence the approximate solution of Sequence Monte-Carlo(SMC)is reached.Meanwhile,simulation experiments are designed in order to test the estimation results.The result shows that the algorithm is able to accurately estimate the conditions of the targets in changing numbers despite the circumstance of disturbing clutter,false alarm and false detection.Thirdly,based on the reality of target tracking,tracing motion control algorithm is designed.According to the Lyapunov function,designing backstepping controller to control the tracking motion.On the basis of this,guidance laws were introduced to optimize tracing path and outputs of controllers.And the result shows that control algorithm is efficient in tracking targets.Finally,build a simulated test system of target tracking.Using six degree of freedom simulated motion platform as the tracking platform,a microwave radar as the sensor,a moving car as the target,to simulate target tracking process.The result shows that algorithm researched in this paper can satisfy the demand of the target tracking mission.