This paper is concerned with the problem of tracking single or multipletargets with multiple non-target specific observations (measurements). For suchfiltering problems with data association uncertainty, a novel feedbackcontrol-based particle filter algorithm is introduced. The algorithm isreferred to as the probabilistic data association-feedback particle filter(PDA-FPF). The proposed filter is shown to represent a generalization to thenonlinear non-Gaussian case of the classical Kalman filter-based probabilisticdata association filter (PDAF). One remarkable conclusion is that the proposedPDA-FPF algorithm retains the innovation error-based feedback structure of theclassical PDAF algorithm, even in the nonlinear non-Gaussian case. Thetheoretical results are illustrated with the aid of numerical examplesmotivated by multiple target tracking applications.
展开▼
