Font Size: a A A

Technology Research On Multiple Radar System Multi-Target Tracking

Posted on:2012-07-06Degree:MasterType:Thesis
Country:ChinaCandidate:X DaiFull Text:PDF
GTID:2218330368977909Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
Data processing of multiple radar targets tracking system is the core component of Navigation Guidance Control. By data processing of radar we will get target trajectories, estimate target movement intention and predict the future position, so as to prepare for appropriate decisions. The background to the dissertation is air target control center, and the research object is the regular movement aircraft during the active flight. By processing relevant trace points and tracks this dissertation gets the comprehensive fusion tracks. After these series of processing targets tracking was achieved.The traditional target tracking algorithm is usually based on Kalman filter estimation theory, Nearest Neighbor Algorithm and Maneuvering target tracking algorithms. With the law of the target motion and the characteristics of the data, this dissertation proposes two programs to extract the target tracks. Program one, the trace points were classified by dynamic partitioning method and K-Means clustering algorithm. Fuzzy function was used as a similarity measure of tracks under different radars. According to the calculation experience set the similarity threshold to justify tracks of the same target;Program two, trace points were searched by wave parameters searching algorithm. This algorithm which can extract the tracks of the same target directly was based on threshold under laws of motion. Time registration of tracks with the same target was carried out by motion model interpolation and cubic spline interpolation separately. Furthermore the accuracy of radar was analyzed and the system error was calibrated. The stability and smoothness of tracks was used as a measurement of radar, and the MSE of points was calculated to express the stability of tracks. After these processing, tracks with the same target were fused by weighting ratio track fusion algorithm.Simulation comparison shows that the tracks extraction by these two programs were both effective. Tracks of the first program are smoother, and the results are more comprehensive, including all the tracks of the targets under radars. The second program is more effective with Uniform and Plane projectile movement, and smaller amount of calculation is the advantage we concerned. After all, tracks of this program exist path delay at the inflection.
Keywords/Search Tags:time registration, target tracking, threshold choice, k-means clustering
PDF Full Text Request
Related items