Robust sensor planning for a partially known moving target: Application to a dynamic X-ray imaging system

This paper discusses the extension of sample-based planning methods to the application of sensor planning for a partially unknown, moving target. This is achieved by modeling the target as a time-varying stochastic process with known mean and variance. Using the unscented transform, the target´s statistics are then propagated through a user-supplied sensing effectiveness metric. The planner is demonstrated for a dynamic X-ray imaging platform tasked with viewing the human knee-joint center during normal, over-ground walking.