The Marginal Distribution Bayesian Filter For Detecting And Tracking The Turn Maneuvering Target

Multiple maneuvering target tracking has been a hot research topic in the academic field.It has a wide application in military and civil fields,therefore is of important research significance.The available research indicates that the existing multi-target Bayesian filter is efficient for multiple maneuvering target tracking.However,it requires that the initial position of the target is known a priori.Due to this requirement,the existing multi-target Bayesian filter fails to detect and track the targets appearing from anywhere in the surveillance region at any time.To solve the problem of target detect and tracking in the2-dimensions and 3-dimensions,two adaptive detecting approaches for the turn maneuvering target are proposed.Applying the proposed detecting approaches to the marginal distribution multi-target Bayesian(MDMB)filter,the MDMB filtering algorithms for detecting and tracking the turn maneuvering target in the 2-dimensions and 3-dimensions are developed,respectively.The main contents of the thesis are as follows:1)We first introduce the multi-target Bayesian filter which is based on the finite set statistics theory(FISST)and its three approximate implementations: the probability hypothesis density(PHD)filter,the cardinality balanced multi-target multi-Bernoulli(CBMeMBer)filter and the marginal distribution multi-target Bayesian filter.Then we outline the traditional sequential processing techniques of track initiation: the rule-based track initiation technique and the logic-based track initiation technique.2)We propose the two adaptive detecting approaches for the turn maneuvering target in the two-dimensions: rule-based detecting approach and logic-based detecting approach.Applying these two adaptive detecting approaches into the MDMB filter,we develop the rule-based MDMB(RB-MDMB)filtering algorithm and the logic-based MDMB(LB-MDMB)filtering algorithm for two-dimensional space.Based on the simulation experimental data,we analyze the tracking performance of the two proposed filters by comparing it with the JMS-GMPHD filter and JMS-CBMeMBer filter and the results show that the two proposed filters in this paper have obvious advantages in target detection and tracking performance..3)We propose the two adaptive approaches for detecting the turn maneuvering target in the 3-dimensions.Applying these two adaptive detecting approaches into the MDMB filter,we then present the RB-MDMB filtering algorithm and LB-MDMB filtering algorithm for three-dimensionas.We also use the simulation experimental data to analyze the tracking performance of the two proposed filters by comparing it with the JMS-GMPHD filter and JMS-CBMeMBer filter and we get the same results as before.