Radar Target Detection And Tracking Based On Metadata

Air surveillance radar system is an important system in both military and civil fields, which plays an important role in pepole's life and country safety. In recent times, military intelligence radar has been faced with many problems which are hard to be sloved, especially, weak target detection and maneauvering target tracking. Air surveillance radar system is a kind of fusion system. It has the same problems as other fusion systems, such as services are not effectively managed, and the perforemance of algortithm need to be improved, and a great number of data can not be effectively managed and shared. Metadata has been discovered through researchs to be the solution to managing and sharing large data effectively. Nowadays, few studies have been done on fusing between metadata and air surveillance radar system. Basing on NSFC, international and military projects, this thesis describes the fusion model between metadata and air surveillance radar system. And the thesis focused on the algorithm of radar target detection and target tracking in radar system. The main points of this thesis are as follows:1) A new model is brought forward, which is a combination of metadata and air surveillance radar system. Service is the base unit in air surveillance radar system.Among the services, the fusion framework between air surveillance radar system and metadata is described. And service is recomposed basing on metadata semantic similarity. Inside service, two fusion methods are introduced namely: Boolean calculation and probability calculation. Through these methods, the improvement of the algorithm performance can be achieved. The metadata-radar system fusing model is composed of fusing framework and fusing methods.2) Aiming at weak target, a new detection algorithm is brought out,which is basing on the combination between the detection metadata with normal detection algorithm by Boolean calculation method. In this thesis, the detection value is lower. Then the detection metadata is set up, which is used to describle the character or processed character of radar target echo. The detection metadata is used to wipe off false plots during the process of the fusion between normal radar target detection result and metadata characters result. As a result, radar system has a better detection probability of faint target. And the detection probability is improved about 10%.3) In order to track crossing flying or maneuvering target, a new kind of data association algorithm is pointed out, which is basing on the combination between plot metadata and normal data association algorithm in probability method. When target is maneauvering, target observation error is processed as target maneauver. So a new kinematic model is set up. According to the new kinematic model, the maneuvering position metadata is computed. In this thesis, the track metadata is used to manage the radar plot and track characters. In order to improve the right probability of the association algorithm, the group of maximum probability is selected. In this thesis, the data obtained from one real flying show is used to test the new algorithm.4) The prototype based on fusing framework between metadata and air surveillance radar system was implemented. The forenamed target detection and tracking algorithms were applied in the prototype. And services composition was completed by service metadata. The prototype adopts the general software in first time, which was realized by using personal computer. The service composition of the prototype can be achieved by the following two steps. The first is to decide the right workflow according to the task. The second was to select the service by integral programming method basing on the semantic similarities of the service metadata. At the end, some simulation experiments are presented.