Research On Multi-UAVs Targets Positioning And Tracking

With the rapid development of UAV,the supervision of UAV in urban environment has become an urgent problem to be solved.At present,the common regulatory means is mainly to locate and track UAV targets.Traditional passive location methods,such as linear iterative methods,need to guess the initial location to obtain the location estimation,and there is a local convergence problem.A two-stage weighted least squares(TSWLS)method is proposed to solve these problems,which uses the time difference of arrival(TDOA)and frequency difference of arrival(FDOA)of signals received by multiple receivers to calculate the algebraic solutions of the position and velocity of moving targets.The method solves the nonlinear estimation problem in two steps by introducing the disturbance variables,and each step is the least squares problem.Compared with traditional passive localization,this method has the advantages of closed solution and low computational complexity.Some traditional correlation algorithms can only associate one measurement with one target,and their computational complexity increases exponentially with the number of targets and measurements in multi-target tracking,while PMHT algorithm allows multiple measurements to be associated with the target,which solves this problem well.It also has good tracking performance for maneuvering target.MM-PMHT algorithm by combining multi-model and PMHT is the focus of this paper.Its main contents are as follows:1.The TSWLS algorithm is introduced,and the nonlinear estimation and weighted matrix solving problem are emphasized.The nonlinear estimation problem is solved by introducing the disturbance variable and dividing the least square estimation into two steps.The optimal weighting matrix is found by minimizing the parameters.2.MM algorithm and PMHT algorithm are introduced.The core of MM algorithm is to give a model set,and then use the model set as many motion models as possible to match the current target motion state,and then filter and fuse all the matched models to get the target state estimation.However,PMHT allows multiple measurements to associate a target,which is easy to expand,and the computational complexity increases linearly with the number of targets and measurements,so it is suitable for multi-maneuvering target tracking scenarios.3.A two-stage weighted least squares method(MP-TSWLS)based on multipath interference is proposed to solve the passive localization problem in multipath interference.This algorithm uses the multipath measurement generated by direct measurement and reflection point to estimate the location of the target,which eliminates the multipath interference and improves the positioning accuracy.At the same time,due to the existence of the added multipath measurement,the problem that the original TSWLS algorithm cannot locate the target when the sensor is coplanar is also solved.Finally,the effectiveness and superiority of MP-TSWLS algorithm are verified by simulation analysis and comparison.4.The MM-PMHT algorithm with Poisson rate is used for UAV target tracking.On the basis of the existing MM-PMHT algorithm,the PMHT algorithm under the mixed model was selected,and the Poisson rate was introduced to characterize the existence of the model.This parameter can be used as an important index for maneuvering model management,so that the redundant maneuvering model can be deleted in real time and the algorithm complexity can be reduced.