Tracking of Multiple Targets Using Interactive Multiple Model and Data Association Filter

Abstract

Tracking of multiple targets using interactive multiple model and data association filter has been developed on PC MA TLAB. The performance of the algorithm is evaluated in terms of estimated and measured tracks in presence of clutter, mode probability, standard deviation, root mean square position error and with varieties of simulated scenarios. Sensitivity of the algorithm to the choice of tracking models and process noise covariance values has been studied. This algorithm has also been evaluated for multi sensor multi target scenario using fusion technique. It was observed that interactive multiple model probabilistic data association filter gives a realistic confidence in the estimates during maneuvers and lower RSSPE during the non-maneuvering phase of the targets.