Generalized PHD filters via a general chain rule

This paper introduces a general chain rule (GCR) for Gâteaux differentials/Gâteaux derivatives, and describes its consequences for multitarget detection and tracking. After describing the GCR and its specific form for functionals and functional derivatives, we use it to derive two new PHD filters: (1) a PHD filter for general models of target-generated measurements with general clutter processes; and (2) a multisensor version of this filter.