Automatic track formation in clutter with a recursive algorithm

Abstract

A recursive algorithm for forming tracks in a cluttered environment is presented. The approach combines the interacting multiple model algorithm with the probabilistic data association filter. The track formation is accomplished by considering two models: one is the true target, with a certain probability of detection P/sub D/; the other is an unobservable target (or no target) with the same model as the former except that P/sub D/=0. The latter represents either a true target outside the sensor coverage or an erroneously hypothesized target. Assuming that the clutter measurements are uniformly distributed, the algorithm yields the true target probability of a track; i.e. it can be called intelligent, since it has a quantitative assessment of whether it has a target in track. The algorithm is useful for low signal-to-noise-ratio situations where the detection threshold has to be set low in order to detect the target, leading to a high rate of false alarms.<>