Research On Target Location And Tracking Algorithms Based On Time Difference Of Arrival

Because of the need of war and civil field,the technology of target location and tracking has always been paid attention by experts and scholars.The passive location and tracking algorithm has become a research hotspot in many countries because the system does not transmit high-power signals,but receives the signals radiated by the target to achieve location and thus has the ability of covert detection.Common positioning methods include passive positioning technology based on Angle Of Arrival(AOA),Time Of Arrival(TOA),Time Difference Of Arrival(TDOA),Frequency Difference Of Arrival(FDOA).This paper focuses on the research of passive location and tracking algorithm based on TDOA.From the point of view of positioning principle,positioning algorithm and positioning accuracy,target positioning technology is studied.The passive tracking technology is studied focusing on the target motion equation,system observation equation and filtering algorithm.Random finite set-based tracking algorithm is a new favorite in the field of multi-target tracking in recent years.Among them,box particle PHD filtering algorithm has been verified in some systems to achieve the same positioning accuracy under the premise of the particle PHD has a great improvement in the speed of operation,this paper also studied this.The specific work is as follows:Firstly,the passive localization algorithm is studied in detail,focusing on the TDOA localization principle,the factors affecting the localization accuracy and several localization algorithms,such as the classical CHAN algorithm with analytic solution,Taylor method and Newton method through iterative operation.A modified Newton localization algorithm with adaptive step size is proposed.Firstly,the traditional highly irregular and non-convex TDOA cost function is approximately reduced to a convex function with regular shape,which can largely avoid falling into local optimal solution.Then the direction of iteration is determined by the modified Newton method to avoid the failure of the traditional Newton method caused by Hessian matrix.Finally,the step size is considered as a one-dimensional variable to optimize and improve the operation efficiency.Then the single target tracking algorithm based on TDOA is studied,and the common motion models such as CV model for uniform moving target,CA model for uniformly accelerated moving target and CT model for constant speed turning are introduced in detail.The essence of TDOA passive tracking is to solve the problem of non-linear filtering.Therefore,this paper studies the extended Kalman and particle filter algorithms.For the maneuvering target whose motion state changes at any time,this paper applies the theory of interactive multi-model to TDOA tracking algorithm,and carries out detailed simulation experiments on the simulation platform of MATLAB.Finally,the box particle PHD filtering algorithm is studied,which essentially applies interval analysis to particle filtering algorithm.After studying the basic principles of operation criteria,inclusion function and shrinkage algorithm in interval analysis,a constraint function for TDOA is proposed to solve the multi-target passive tracking problem using box particle PHD algorithm.