Data Association Algorithm In Dense Target Environment

Data Association is the pivotal part of the multi-target tracking system, as well as the problem in the field of target tracking. The environment of tracking is complicated and multivariate at present and science and technology develop rapidly, field of military or civilian areas put forward the ever-increasing demands of data association, such as air warning, battlefield surveillance, air and water traffic control and navigation systems, etc. The present challenge of the data association is the large computation, the bad performance of real-time and that the algorithms can’t track the closed moving targets effectively. Therefore, the article mainly studies the data association algorithm in dense target environment, and the major work is summarized as follows:Firstly, the paper introduces the basic theory of multi-target tracking and data association, focuses on analysis and comparison of the nearest neighbor algorithm, probabilistic data association and the joint probabilistic data association, and analyzes the influence of several ways of obtaining the clutter density to the probabilistic data association algorithm as well as the modified probabilistic data association algorithm based on distance.Then, because of the limitation that the joint probabilistic data association can’t be applied to the practical engineering, the paper reduces the time of data association from decreasing the confirm matrix and dealing with the public measurements. From the aspect of decreasing the confirm matrix, the paper proposes the joint probabilistic data association algorithm based on grid and connection; from the aspect of dealing with the public measurements, the paper quotes the algorithm that assigns the public measurements to the related the targets according to certain probabilities, so that the probabilistic data association can track multi-target.At last, the paper quotes the concept of random sets to the joint probabilistic data association algorithm to solve the problem that the traditional data association algorithms can’t distinguish the closed moving targets. Through targets in the random sets switch orderly, it implements the tracking to the closed moving targets.