Multiple hypothesis track-to-truth assignment with a minimum sojourn time

Multitarget-Multisensor tracking metrics have been studied extensively in recent years. These tracking metrics require a unique assignment of tracks to truth objects at each scoring time. There are two typical approaches to address this track-to-truth assignment problem. The first approach involves developing content vectors for the measurements (i.e., vectors indicating the percentage of each truth object content in a measurement) and tracing through the computer simulations and tracking algorithms the contribution of each truth object to the tracks via the content vectors. The second approach involves the kinematic assignment of tracks to the truth objects. This second approach does not require any special design considerations in the computer simulation, and it can be used with flight test data. In this paper, a multiple hypothesis or multiple frame approach is taken to the kinematic assignment of tracks to truth, and a minimum constraint is placed on the sojourn time of those assignments to eliminate artificial switches when tracks on closely spaced and possibly unresolved objects are scored. The multiple frame approach to track-to-truth assignment is applied to simulated data and the results are compared to assignment results from single frame assignment with no constraint on the sojourn time and content vectors.