Study On Maneuvering Target Tracking And Multi-target Association Algorithm

The study of target tracking theory and data association algorithm all have far-reaching significance not only in the military field but in the civilian areas. With the continuous development of technology, a variety of new techniques being applied to target tracking theory, but at the same time, the application environment is increasingly complex, how to improve the performance of target tracking algorithms and the data association algorithm quickly have become serious problems. For the filtering algorithms and the data association algorithm are the core and difficulty of maneuvering target tracking theory, this thesis focus on the two aspects and conducted in-depth research.Firstly, the basic principles and components of the theory of maneuvering target tracking and were introduced, and some maneuvering target motion models and non-linear filters have been discussed in the next section, at last, two improved algorithms based on strong tracking filtering have been proposed. The improved algorithm can not only ensure the features of the original filters but can also improve the robustness of the original filters. The abilities to respond to mutations of the two filters have been improved. At last, two kinds of simulation environment have been given to verify the validity of these two algorithms.Secondly, the characteristics and tasks of the maneuvering target tracking were studied in detail and a brief description of some classical data association algorithms of multi-target was given. A target has a fuse of uniformly accelerated motion and uniform motion has been applied for the analysis of data association algorithm for single-target. Simulation results show that probabilistic data association algorithm based on interacting multiple models is superior to the probabilistic data association algorithm when the target’s movement can’t be described only by a single model. Two targets with cross-track in the two-dimensional plane were used for multi-target data association algorithm. Simulation results show that the tracking error of GPDA is less than the tracking error of JPDA.