earticle

영어

In distributed sensor networks, track association and track fusion become difficult due to the existence of various uncertainties in multiple target tracking (MTT). In an actual tracking system, state estimates of a local track are usually transmitted from local nodes to the global node by message, and each message generally contains single state estimate. Based on this fact, one can define two state estimates of a local track in continuous times as a tracklet. Then, local track-global track association can be divided into tracklet-global track (T2GT) association in real time. Hence, a T2GT association method based on Hough transform (HT-T2GT) is proposed. By Hough transform, all the tracklets in the same interval can be mapped into a set of points in Hough space, and the track association problem can be transformed as one of point clustering in Hough space. The maximum entropy fuzzy c-mean (ME-FCM) clustering method is used to realize T2GT association. In addition, a T2GT fusion method based on the support degree function (SDF-T2GT) is developed for track fusion. The experimental results illustrate that the proposed methods can respectively realize T2GT association and track fusion in the situations with multiple local nodes, reduce the average time of updating global tracks and satisfy the requirement of real-time processing in the global node. It achieves higher association processing rate than other two track association methods.