Joint tracking and classification of extended object based on support functions

This study considers joint tracking and classification (JTC) of an extended object using measurements of down-range and cross-range extent. Using such measurements, existing approaches handle only tracking, that is estimating the kinematic state and the extension. In many practical applications, tracking and classification (e.g. classifying the object by its size and shape) are highly coupled (i.e. they affect each other) but are handled separately. For JTC of extended objects, this study deals with this problem jointly by integrating class-related extension information (i.e. the size and shape characteristics distinguishing objects of different classes) into a support function model. This facilitates the derivation of their JTC algorithm for jointly estimating the kinematic state and object extension and obtaining the probabilities of the object classes. In the proposed JTC algorithm, the useful information between the tracker and the classifier is sufficiently exchanged to improve overall performance. Furthermore, they also propose an effective method to fuse object extension estimates. The benefit of what they proposed is illustrated by simulation results.