Sensor Networks by Gianluigi Ferrari

Author:Gianluigi Ferrari
Language: eng
Format: epub
Publisher: Springer Berlin Heidelberg, Berlin, Heidelberg

(36)

where S is the track set , is nn x × n x matrix and n x is the dimension of the state vector. Matrices and are given by (37) and (38), respectively,

(37)

(38)

The covariance matrix of the fused track is given by

(39)

7.3 Tracklet Fusion

As tracklets are track data not cross-correlated with the common information among the platforms, for the data fusion purposes they can be treated like measurements and be associated to tracks of other platforms. That solves the problem of data synchronization typical to track-to-track fusion considered above. Since not all the correlation among the versions of the same track maintained by several platforms can be removed, such an approach is only reliable when dealing with targets having small maneuvering index. There were proposed a number of methods for calculating tracklets [40, 41]. In this chapter, we use an algorithm in which a tracklet is the state of a track decorrelated with the state of the same target at the time when the last tracklet was transmitted. For this decorrelation operation the older state requires to be predicted to the time of the more recent one. For tracklet computation, the state of the corresponding track must be observable from the measurements received after the last communication of a tracklet corresponding to the track. That is, at least two measurements are required if the target state vector contains the position and the velocity. Assuming that the last tracklet was transmitted at the time k, the track was last updated at time step k+i, and there are enough measurements between time step k and time step k+n, the tracklet and the corresponding covariance matrix at time step k+n are given according to [33] by (40) and (41), respectively. That is,

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.