A multi-sensor generalized labeled multi-Bernoulli filter via extended association map

The generalized labeled multi-Bernoulli filter, or Vo-Vo filter, is the optimal Bayes solution to the multi-target tracking problem. Conceptually extension of this solution to the multi-sensor case is straightforward. However, implementation of the multi-sensor generalized labeled multi-Bernoulli filter is nontrivial. In this paper, we discuss a number of implementation strategies for the multi-sensor generalized labeled multi-Bernoulli filter. By extending the association map for single sensor, we propose an extended association map and describe an efficient implementation of the multi-sensor generalized labeled multi-Bernoulli update.