Angle-Only Tracking of Manoeuvring Targets using Double Sets of Multiple Range Models and Probability Mass Diffusion

This paper presents a solution to the problem of tracking of manoeuvring targets by angle-only measurements from a single, manoeuvrable platform. The solution combines the properties of accurate range estimates of non-manoeuvring targets with the ability to track manoeuvring targets in an all recursive setting. While previous methods use combinations of recursive and batch processing and/or threshold type manoeuvre detection, the proposed solution has these properties inherently, with a "soft" manoeuvre detection mechanism. The solution, in this paper, is based on a combination of two previously presented ideas, the Multiple Range Models and the recently presented Probability Mass Diffusion filter. In the paper the general problem of angle-only tracking is discussed and the basics of the proposed solution are given. Performance is shown by Monte Carlo simulations of a wide range of target scenarios.