Explore chapters and articles related to this topic
Application Tracking and Navigation
Published in Bin Jia, Ming Xin, Grid-based Nonlinear Estimation and Its Applications, 2019
Note that the PDAF does not include the initialization algorithms. The clutter is assumed uniformly distributed in the measurement space and the uncertainty of the state is represented by the Gaussian distribution. The PDAF treats all but one measure in the validation region as clutter. It suffers the track coalescence problem. Different from the PDAF, the joint probabilistic data association filter (JPDAF) treats the interacting targets as a cluster. The JPDAF calculates the association probability for all measurements and the active targets in the cluster. Hence, one measurement is used to update more tracks if the measurement is within more than one validation gates.
Introduction to Multitarget Tracking
Published in K. V. Ramachandra, Kalman Filtering Techniques for Radar Tracking, 2018
The JPDAF [1–8,17] is identical to PDAF except that association probabilities are now computed using all observations and all tracks. The state estimates, gain, and covariance of JPDAF are computed using (10.14) to (10.20). The probability computations of (10.23) and (10.24) are now extended to include multiple tracks.
Boosting the multiple aircraft online tracking performance via enriching the associated data with fused targets features
Published in International Journal of Image and Data Fusion, 2023
A. M. Awed, Ali Maher, Mohammed A. H. Abozied, Yehia Z. Elhalwagy
To solve the data association problem there is main question in assigning detection to target is which detection should be assigned to the target or deleted if it was false or added to a new detector. In general, classical data association approaches are used like the Joint Probabilistic Data Association Filter (JPDAF) (Fortmann et al. 1983) and Multiple Hypotheses Tracking (MHT) (Reid 1979) they consider all possible associations between targets and detection responses. Alternatively, the Hungarian algorithm (Kuhn 1955) and the greedy search algorithm (Yang et al. 2009) can be used to recursively choose the best assignment between a set of detections and the set of targets as in Figure 4, recently, tracking by tracklets approaches were exploited (Wang et al. 2014, Yang and Nevatia 2014, Zhang et al. 2015). These techniques are deal with data association process as a set of local trajectory fragments.