The identity management problem — A short survey

The identity management problem is the problem of probabilistically keeping track of the association between target tracks and target identities, based on observations made by sensors. Updates of the belief state can happen because of new sensor observations reflecting on target identity, or because targets come near each other so that their identities become confused or mixed. Since the space of all possible associations grows factorially with the number of targets, it becomes important to find compact representations for distributions over associations and efficient implementations of the associated filter operations. In this note we introduce and describe the identity management problem and then place in context and briefly survey some the earlier work on this problem by us and others.