Data Association Research About Targets Among Ship Formation Based On Heterogeneous Spaceborne Reconnaissance Sensors

Because of the superiority of using spaceborne optical imaging reconnaissance and spaceborne electronic reconnaissance to monitor the targets, and the importance to surveillance the ship formation, using spaceborne reconnaissance instrumentality to monitor the ship formation is helpful to located and identify the ships among the ship formation and grasp the mastery of the information. The precondition of located and identify the targets is to find out the relationship between the targets and the measurement, this called data association. Based on this background, the main research of this dissertation is the data association algorithm in the situation that the location precision of different sensors are entirely different that can not using the traditional association algorithm. This dissertation made the following findings.In order to using the target location information with great difference in precision, based on the trait of formation movement, define the similarity measure based on the formation structural characteristic to solve the data association problem; In the situation that the target attribute information obtain by heterogeneous sensors are in different feature space, construct the similarity measure to solve the data association problem by using the attribute information; Because the targets attribute information and the targets location information are totally different, in order to improve the association performance, put forward the similarity measure by using those information; When we obtained the sequence data, we can use the Kalman filter in data association. In this dissertation we use the formation structural characteristic and the target attribute information to update the Kalman filter input value to improve the data association performance. In this article, based on the real situation and the characteristic of the actual problem, construct different scenarios in simulation, the result shows the effective of the algorithms.