The Research On Radar Maneuvering Target Tracking Filter Algorithm

Target tracking has a wide application in both military and civilian fields, so much more attention has been paid on the research of it, especially the problem of maneuvering target tracking. The challenge we faced in the target tracking is its two discrete uncertainties: the measurement uncertainty and the target moving mode uncertainty. A lot of deep and systematic research on maritime radar's maneuvering target tracking research has been done on the basis of studying other people's research results in order to track target accurately, efficiently and stably.Firstly, the principle of target tracking was summarized including target tracking mode and some important filtering methods focusing the research on Kalman filter and its changing mode in nonlinear system. Secondly, the research on the maneuver detection and identification algorithm has been made. The interactive multiple mode and another one base on fuzzy inference system has also been studied. Finally, simplified filter methods has been studied for a better implication. The main points of the paper are as follows:The EKF and Bar-Shalom's CMKF-D algorithm in a nonlinear system has been studied. It is proved that both of them are effective when the measurement is obtained in a nonlinear system by the simulation.The research of maneuvering target tracking has been made and a new algorithm with a better implication was proposed on the basis of studying the current statistic mode. An adaptive filter based on the analysis of error vector has also been studied. It was base on CV model and the adjustment strategy of state error covariance matrix is very simple, which would simplify the maneuvering target tracking.The interaction multiple mode algorithm and a fuzzy interaction multiple mode algorithm has been analyzed. The fuzzy interaction multiple mode algorithm obtains the matched degree of each filter model in the designed model set based on fuzzy inference system, by which the model probability in the existing interactive multiple model algorithm is replaced and the calculation complexity is decreased obviously, which is better for its usage in real time target tracking.The simplified Kalman filter has been studied for its implication. An adaptive sub-cycle mechanism to solve the problem when is the time to start a new Kalman gain calculating progress has been proposed in the studying of sub-cycle Kalman filter. In this way the calculation complexity is decreased a lot, which is good for the implication of maneuvering target tracking.