A Fuzzy Logic Framework to Estimate the Angular Turn Rate for High-Performance Target Tracking

Most tracking algorithms are model based because knowledge of target motion is available. A successful target tracking depends on the choice of a good model of the target; which facilitates the extraction of useful information about the target´s state from observations. A good model-based tracking algorithm will outperform any model-free tracking algorithm to a great extent if the underlying model turns out to be a good one. In this paper, we propose a fuzzy logic framework to expect the angular turn rate and, thus, the appropriate model for target tracking using the particle filter. The coordinate turn (CT) model with unknown turn rate is used to calculate the expected angular turn rate. The fuzzy logic system is comprised of double-input single-output; which is presented by fuzzy relational equations. A canonical-rule based form is used to express each of these fuzzy relational equations. The dynamics of the high-performance target are modeled by multiple switching (jump Markov) systems. The proposed algorithm showed better performance in tracking a maneuvering target in relative slow scan periods compared to fuzzy integrated multiple models (FIMM).