Interacting multiple model probabilistic data association - en studie av målfølgingsalgoritme for aktiv sonar

Abstract

This document conducts a study of a target tracking algorithm, based on interacting multiple models in combination with probabilistic data association, derived, presented and discussed in the UNISON 2005 project. The effects on the validation region and the track properties, by adjusting the parameter settings, are studied. A proximity filter is derived, presented and discussed. Questioned is the use of inverted range as a state. This is derived from systems using passive sonars, but the algorithm is for a system based upon using active sonars.