engleski

Tracking Multiple Moving Objects Using Adaptive Sample-based Joint Probabilistic Data Association Filter

In this paper we present a probabilistic method for tracking multiple moving objects. Joint probabilistic data association filter is used for assignments between detected features and objects being tracked. A particle filter is used for representation of underlying object state uncertainty. Novelty of our approach is particle number adaptation. Experiments done on real world and simulated laser range data show that our algorithm is robust and accurate in tracking multiple objects.

jpdaf; particle filter; kld-sampling; tracking; moving objects

nije evidentirano