Multiple Target Tracking Of Cardinalized Probability Hypothesis Density Filter With Unknown Model Parameters

In recent years,more and more experts engaged in the study of multi-target tracking technology based on random finite set(RFS).Multiple target tracking has also been aplied to many fields,such as military fileds and civilian fields.But the target tracking environment becomes increasingly complex with the development of science and technology.In order to adapt to such a tracking environment,a variety of multi-target application systems for multi-target tracking accuracy and performance of the higher requirements,which for the original target tracking theory and method is a greater challenge.Therefore,it is significant to study the Multi-target tracking technology more deeply.During the multi-target tracking process,the performance of the multiple target tracking algorithm and the tracking accuracy are influenced by many factors,such as the uncertainty of measurements and goals,which assumes that certain model parameter is known a priori,but the real values of these parameters are not consistent with the assumptions in the actual tracking environment.This can lead to a goal of error or failure to follow.So this paper makes a simple summarization of the adaptive Multi-target Tracking algorithm,and introduces the filtering algorithm of the cardinality probability hypothesis density(CPHD)filter with unknown clutter,unknown detection probability,unknown new target intensity and unknown target signal-to-noise ratio.Among them,it is an important task to improve the performance of multi-target tracking algorithm to estimate the detection probability,the number of clutter and the spatial distribution of adaptive target in the process of tracking target.On the base of Mahler proposed a CPHD filter which can adapt itself to the clutter rate and the target detection probability.The CPHD filter for the unknown clutter and the detection probability have multiple complex integrals that cannot be solved by the closed analytic solution,this article adopted gaussian mixture(GM)to give the closed form of filtering recursive.This paper introduces the filtering process and coding process of the Gaussian mixture of the two filters in detail,and shows the MATLAB simulation experiment in the end of third chapter and the fourth chapter.