1 Introduction
Tracking a moving ground target from a unmanned aerial vehicle (UAV) is a key requirement for many information, surveillance, target acquisition, and reconnaissance (ISTAR) [1] applications, which is one of the many uses of UAVs. In target tracking, the observability of target is achieved by adequate maneuvers of the sensor mounted on the UAV. Since the sensing capabilities of UAVs are constrained by the UAV dynamics and the onboard sensor performances, one of the major research challenges is UAV action decision-making, which means how to plan ahead for the sensor platform movement to maximize target localization accuracy at a future time.