Researches On Maneuvering Target Modeling And Tracking Method

With the rapid development of the technol ogy and science, the environment of the modern war has been more and more com plex. The m aneuverability of tar gets is becoming more and more com plicated and inconstant; meanwhile the demand of the tracking performance index also is increasing. Hence, the target tracking technology becomes the research hinge in the military and civil. In the field of maneuvering target tracking, nonlinear filtering technology is the main tool, the modeling of the maneuvering target is the essential for the maneuvering target tracking, and the m odel of structural is an im portant means of implementation of the m aneuvering target tracking. Therefore, this thesis studies fo r the m aneuvering target tracking from the three statements above.First of all, this th esis illustrates the basic p rinciple of the maneuvering target tracking, and analyzes the existence of nonlin ear factors in the contem porary tracking system. And three nonlinear kalman filter algorithms are analyzed in detail. Including Extended Kalman Filter algorithm, Unscented Kalman Filter algorithm and Cubature Kalman Filter algorithm, and then through the simulation of th e maneuvering target tracking, the advantages and disadvantages of the three no nlinear filter algorithms are analyzed. And the three algorithms are applied in the subsequent simulations.Then, this thesis intro duces several maneuvering target motion m athematical model, and expounds the main principle of the current statistical m odel and Jerk model in detail. Through the experimental simulation, this thesis analyzes the current statistical model and Jerk m odel for the effect of the maneuvering target tracking. Analyzed the current statistical model, the problem of insufficient tracking precision in the non-motor segment, and based on this, expounds the current statistical m odel based on fuzzy adaptive, compared with the current statisti cal model for the simulation of the original and improved the current statistical model is verified in m otor and non-motor vehicle segment than orig inal current statistical model for maneuvering target has b etter tracking precision.For the res earch of th e model structural, this thes is first researches on the interactive model algorithm, which is the m ost widely applied algorithm in the f ixed structure multiple model. Then th e interactive model algorithms of dif ferent model number were compared, which led to the lim itations of fixed structure multiple model algorithm. Then the basic principle of the variable structure multiple model algorithms and the cur rent four of the m ain variable structure multiple model algorithms are expounded, including the Model-G roup Switching algorithm, the Likely-Model Set algorithm, the Adaptiv e-Grid algorithm and the Expecta tion-Mode Augmentation algorithm. And the af ter three algorithms with interactive multiple model algorithm of the fixed structure are simulated. For the pr oblem of insufficient maneuvering tracking ability in the orig inal Expectation-Mode Augmentation algorithm, this the sis puts forward the Expectation-Mode Augmentation algorithm based on Strong Tracking Filter, which further improves the accuracy of th e expectation mode expansion algorithm in the motor segment. It expands the thinking of the Expectation-Mode Augmentation into the Likely-Model Set algorithm, this thesis puts forward the improved Likely-Model Set algorithm which is introduced the thinking of the Expectation-Mode Augm entation, and through the simulation of comparing with interactive multiple model algorithm of the fixed structure, confirms that the improved algorithm not only improves the tracking precision, but also improves the efficiency of the algorithm by using fewer models.