Comparison of two augmented PDA filters

Probabilistic data association (PDA) is a technique for performing data association in tracking applications where the presence of false and missing data causes measurement origin uncertainties. In some applications, additional feature parameters are available with the measurements. In this paper, two techniques for incorporating this “augmented data” into the filter are discussed and compared. The new technique developed is the augmented state PDA filter. The second approach is the augmented PDA with feature measurements. Implementation advantages of the former filter are described, and analysis and simulation results are given to show that for certain models the two filters have comparable performance