It can be said that multi target (single sensor) algorithms are consisted of two phases:
1 - data association
2 - single target tracking for each target
Now the question; is it possible to use one of the existing estimation methods designed to handle measurement delays as a filtering mechanism in phase 2?