Saccades planning with kinetic TSP for distant targets identification

This paper considers the problem of designing an active observer to plan a sequence of decisions regarding what target to look at, through a foveal-sensing action. We propose a framework in which a pan/tilt/zoom camera schedules saccades in order to acquire high resolution images (at least one) of as many moving targets as possible before they leaving the scene. An intelligent choice of the order of sensing the targets can significantly reduce the total dead-time waste by the active camera and, consequently, its cycle time. We cast the whole problem as a particular kind of discrete optimization. In particular we will show that the problem can be solved by modeling the attentional gaze control with a novel form of kinetic traveling salesperson problem (KTSP), reformulated as a time dependent orienteering (TDO) problem.