Multi-source Relevant Surveillance Radar Data Fusion Research

With the airborne navigation system, the ADS-B system can provide more exactinformation about the aircrafts than the traditional RADAR system. As the newgeneration of the Air Traffic Control technology, ADS-B is promoted by theInternational Civil Aviation Organization(ICAO), and it will play a more important rolein the future Air Traffic Control system. The implementation of ADS-B system isdifferent from the traditional RADAR system. Without the interrogation from groundstations, the airborne transmitter sends messages of the aircraft automatically. Alongwith the increasing of aircrafts, the flight route is becoming busier. So the increase ofroute capacity is becoming more exigent. With ADS-B, we can improve the precision ofsurveillance and shorten the distance between two aircrafts, which can help to improvethe capacity of flight route. So, along with the widespread use of ADS-B technology, itwill be more and more common that two means of surveillance work together duringthe flight. Hence, the investigation of how to efficiently make use of two types ofinformation has realistic significance.This thesis focused on how to efficiently make use of the ADS-B information andSSR information, in order to get more exact information about the aircrafts. The maincontent of this thesis is listed as following.First, this thesis introduced the development of ATC system and data fusion. Afterthat, it introduced the ways of SSR and ADS-B worked, and the content of the twotypes of messages.Then, the thesis analyzed the flow of data fusion. And the observation noise ofADS-B and radar tracks were simulated in this thesis, it also compared the effect ofkalman filter and IMM filter.In this thesis, we used the double threshold correlation algorithm which based onthe Nearest Neighbor (NN) algorithm, and it cost less simulation time of the correlation.The algorithm executed the Nearest Neighbor algorithm at first, then, the tracks whichpassed the NN algorithm’s verification can be taken to the double threshold correlation’s verification. When more than two tracks correlated to one target, the ambiguityprocessing was done. Next, the correlated tracks were weighted fused depending onwhat the NACp were. The simulation results show that the fused tracks’ precision wasmore accurate than the single radar tracks.At last, we divided the fusion system into two cases, system initialization andsystem update. Both cases were carefully analyzed. We proposed system initializationalgorithm and system update algorithm for both cases. Then we simulated corres–ponding environment for both cases and both algorithms were tested.