Three-dimensional target motion analysis using angle-only measurements

Infrared search and track (IRST) system is an important passive tracking system widely applied in military and civil domain. Locating the moving target by the angle-only measurements is difficult because of nonlinearity and variable observability. This paper studies the angle-only tracking (AOT) in three-dimensional (3D) state space. We discuss the effective AOT approaches using the Gaussian mixture probability hypothesis density (GM-PHD) filter and the Gaussian mixture cardinalized probability hypothesis density (GM-CPHD) filter. Moreover, by constructing different multi-target simulation scenario, the tracking performances of PHD and CPHD are compared. And CPHD has better estimation evaluated in terms of the optimal sub-pattern assignment OSPA metric, but with low computing efficiency and slow response to the target number change.