State Estimation of Objects that Have and Have Not Been Observed
Lance A. Page, William Hall · 2003
This paper describes an estimator called the Information Model that computes three types of information: tracks containing state estimates for individual objects that have been observed; geographic distributions of undetected objects that have not yet been directly observed but whose presence is suspected; and the geographic distributions of total object counts, including both tracks and undetected objects. Collectively, these outputs support a variety of decision-making processes about finding objects, monitoring objects, doing things to objects, or simply avoiding objects. As its inputs the Information Model accepts measurements from a broad class of sensors whose observations can be packaged as contacts, for example, Search Radar, Electro-optical, and Synthetic Aperture Radar sensors. The Information Model can integrate observations with widely varying geographic coverage, probabilities of detection, false detection rates, and update frequencies. It can also use additional, independent expectations of geographic distributions of the objects present. This paper defines the estimator’s inputs and outputs, describes its computations, and presents results derived from simulations.