Data Fusion In Radar Networking System

In C4ISR systems, informations provided by single sensor can no longer match needs of battlefield, while sensor-networks, as an effective way to enlarge sense-scope, are payed a great attention. A certain sensor-network fence system is a key project of national 10th five-year plans. The research and development of the system is pretty meaningful to keep up with the tide of this technology.In sensor-network systems, data fusion is the key. It is a organic whole of a series of technologies, such as target detection, track association, track filtering and so on.In the first chapter, the author introduce the research background, developments, data fusion and peculiarity of the systems which are combined with different kind of sensors and aim at low-space fast targets.In the second chapter, related technologies are presented. These are algorithms for track-filtering, filtering on condition of multisensor, data association on condition of multitarget and so on.The last three chapters is the key-point of the whole paper, one of the author's three contributions is discussed in each.The third chapter is about the design of the center-stage fusion process. The author adapted normal methods for the system.The fourth chapter is about kalman trackers and an improvement. The author replace the old polynomial motion models like CV, CA with a new nonlinear motion model witch have constant upright and parallel accelerations, and gain a much better performance on curvilinear flight paths.The fifth chapter is about a space-transformation method which considers the bend of the earth's surface. The author first demonstrate it is necessary to use ellipsoidal model which is more precise but a little bit complicated, then he educe the transform formular. This research is useful for all large sensor-networks.