Generalized Correlation of Multi-Target Track Data

Abstract
A general expression based upon a maximum likelihood approach is derived for the optimal association and correlation of observations into tracks. Main features of the approach are the definition of a score function and a unified approach to track initiation, confirmation, gating and deletion logic. Simplifications are made to derive rules for suboptimal sequential processing and this sequential method is shown through simulation to give improved performance wver a more standard approach. Further applications are indicated.