Multi-target particle filtering using MUSIC as a pseudo-likelihood

  • 1D_2targets_3clusters.avi: This file corresponds to Fig. 1 in the paper and illustrates the MUSIC based multi-target PF implementation in one dimension.

  • 1D_2targets_4clusters.avi: This file illustrates the MUSIC based multi-target PF implementation in one dimension. The specifications in the caption of Fig. 1 in the paper are used, except that here, 4 clusters are formed.

  • 1D_3target_4clusters.avi: This file illustrates the MUSIC based multi-target PF implementation in one dimension. The specifications in the caption of Fig. 1 in the paper are used, except that here, 4 clusters are formed and 3 targets appear in the scene.

  • DataAssociation1.avi: This file corresponds to Fig. 4 in the paper and illustrates the ability/inability of the various filters to overcome the data association problem. The two dimensions of the multi-dimensional PF are plotted on the same axis.

  • DataAssociation2.avi: This file corresponds to Fig. 4 in the paper and illustrates the ability/inability of the various filters to overcome the data association problem. The multi-dimensional PF is plotted in two dimensions.

  • ParticleHijack.avi: This file corresponds to Fig. 3 in the paper and illustrates the ability/inability of the various filters to overcome the particle hijack problem.

  • ResolvingCloseTargets.avi: This file corresponds to Fig. 5 in the paper and illustrates the ability/ inability of the various filters to resolve close and crossing targets.