Research On Sensor Control Method Based On Gaussian Mixture Multi-target Filter

Multi-target tracking(MTT)technology has always been a hot topic in engineering application and theoretical research.In essence,it is a time-varying joint decision and estimation process.In MTT system,in order to maximize the quality of the received measurement information,a sensor is controlled through a reasonable optimization decision-making process.Thus,it can improve the overall level of information fusion.At the same time,it is of great practical significance for solving MTT in complex environment.In recent years,MTT algorithm based on finite set statistics(FISST)has been widely concerned,because it can avoid the complex data association.This method system not only has a solid mathematical foundation,but also provides great convenience for the research of sensor control in MTT.In view of this,on the basis of FISST,the corresponding sensor control strategies via Gaussian mixture multi-target filter in MTT are mainly studied in this thesis,and the main contents are as follows:1)In consideration of the sensor control problem for MTT,the corresponding sensor control strategies are studied via Gaussian mixture probability hypothesis density filter(GM-PHDF)with FISST.First,on the basis of the Cubature Kalman GM-PHDF(CK-GMPHDF),the information gain between multi-target probability densities is given through the Bhattacharyya distance between the two Gaussian distributions.Then,taking advantage of this information distance,this thesis proposes a corresponding sensor control strategy.Furthermore,a joint sampling method of Gaussian particle is designed to sample the predicted Gaussian component of multi-target filter.Subsequently,a set of weighted particles are used to approximate the multi-target statistical characteristic,and their weights are updated with the ideal measurement set.Next,a Rényi divergence based sensor control strategy which has better adaptability is proposed.Finally,a detailed Gaussian mixture implementation of the posterior expected number of targets(PENT)is given.Simulation results verify the effectiveness of these algorithms.2)In order to further expand the sensor control based on Gaussian mixture probability hypothesis density filter(GM-PHDF),this thesis studies Cauchy-Schwarz(CS)divergence based sensor control via GM-PHDF.Via the approximate statistical characteristics of multi-target probability density,the PHD expression of CS divergence is derived.On the basis of the CK-GMPHDF,the proposed control method,which is based on the criteria of maximizing the information gain,is given under information theory where the rewardfunction is calculated by the CS divergence between the multi-target prior and the multi-target posterior densities.Simulation results verify the effectiveness of the proposed method.3)At last,this thesis proposes the sensor control strategy using Gaussian mixture multi-Bernoulli filter based on the finite set statistics(FISST)theory.First,this thesis gives the implementation of the Cubature Kalman Gaussian mixture cardinality balanced multi-target multi-Bernoulli filter(CK-GMCBMeMBerF),and extracts the Gaussian mixture component to approximate multi-Bernoulli density.In addition,we study the solution of the Cauchy-Schwarz divergence between the two Gaussian mixture distributions,and derive the information gain corresponding to the change of multi-target probability density.Then,the corresponding sensor control method is proposed.Moreover,a detailed GM implementation of the posterior expected number of targets(PENT)criteria is given based on CK-GMCBMeMBerF,and the corresponding sensor control strategy is studied with GM-PENT as the evaluation criterion.Finally,simulation results verify the effectiveness of these proposed algorithms.