Abstract
The multiple target tracking (MTT) performance of a new combination of the fuzzy interacting multiple model (FIMM) algorithm and the probabilistic data association filter (PDAF) is investigated. The ability of a set of these FIMMPDAFs to maintain the tracks of multiple targets in a cluttered adaptive cruise control (ACC) environment is compared to that of the likelihood approach, the IMMPDAF. The differences between the two methods are highlighted and simulation results for a typical highway driving scenario demonstrate the performance of each approach. These results show that both the IMMPDAF and FIMMPDAF strategies are capable of tracking multiple vehicles with low RMS position errors, while the FIMMPDAF appears to detect the initiation of a target maneuver more rapidly by adjusting model probabilities more quickly.
Distributed multiple model joint probabilistic data association with Gibbs sampling-aided implementation
