Research On Low-altitude Maneuvering Target Tracking Method

With the rapid development of science and technologies,the number of low-altitude targets is increasing,which poses a great threat to public safety and national defense security.Therefore,it is crucial and imperative to strengthen the research on low-altitude target tracking.Low-altitude maneuvering targets have the characteristics of low flight altitude,fast speed,and high maneuverability.In addition,ground clutter,sea clutter,and ground object interference often cause high false alarms,so many sensors have poor tracking accuracy.Therefore,to timely and accurately track low altitude maneuvering targets,the interactive multi-model（IMM）,data association,and filtering algorithms are used to study the low-altitude target tracking algorithm in this thesis.The main research contents include:Firstly,aiming at the problems of low tracking accuracy and easy divergence of the traditional IMM algorithm in the clutter background and complex maneuvering conditions,IMM is combined with the probabilistic data association algorithm（PDA）.Based on the framework of integrated IMM-PDA,an improved IMM-PDA low-altitude maneuvering target tracking algorithm based on the adaptive expanding tracking gate is proposed.On the one hand,the IMM algorithm of transfer probability correction based on the current statistical model is proposed for the complex maneuvering of low altitude targets.The algorithm firstly uses the probability difference between two consecutive time points in the IMM model to adjust the probability transition matrix in real time,which solves the problem of low tracking accuracy caused by the fixed transfer probability matrix of the IMM model.Then the adaptive current statistical model is incorporated into the IMM model set to improve the accuracy and adaptability.For another,an adaptive extended tracking gate is designed.The size of the elliptical gate is adjusted by the threshold constant.The results show that this method can effectively improve the accuracy of target tracking estimation in a clutter background and has practical value.Secondly,to address issues of poor real-time performance and particle dilution in unscented particle filter,a maneuvering target tracking algorithm based on adaptive unscented particle filter is proposed.The algorithm first uses the minimum slope simplex unscented transform to reduce the number of selected sigma points,and uses the scale unscented transform to reduce the local effect.Then,a new fading factor is derived based on the strong tracking theory.The factor is introduced into the state prediction error covariance matrix and gain,so that the algorithm can perform maneuvering detection and effectively balance the contribution of predicted and measured values.In order to solve the problem of particle impoverishment,the particle weight is optimized by introducing a weight adjustment factor to improve the proportion of particle diversity.On this basis,Kullback-Lerbler distance sampling is integrated to adaptively adjust the particle size,so as to improve the computational efficiency from the perspective of reducing the number of particles.The simulation results show that this algorithm has satisfactory performance and achieves the purpose of improving the tracking accuracy and computational efficiency of the unscented particle filter.