An Improved Probabilistic Data Association Method for Multi-Target Tracking in Clutter

yazhao wang,yingmin jia,junping du,fashan yu
DOI: https://doi.org/10.1007/978-1-4471-2467-2_164
2012-01-01
Abstract:Improved data association methods for multi-target tracking are discussed in this paper. We introduce an exquisite gating technique based on the well-known probabilistic data association (PDA) filter, which calculates the association probabilities to the targets being tracked for each validated measurement. An important feature of the new algorithm is that it can divide the joint validation region into several parts in which the association probabilities can be calculated with more accuracy. The algorithm is tested and compared with both the generic PDA and joint probabilistic data association (JPDA) filter. Simulations are also presented to illustrate the effectiveness of the algorithm.
What problem does this paper attempt to address?