An entropic criterion for minimum uncertainty sensing in recognition and localization. I. Theoretical and conceptual aspects

A criterion is presented for the automatic selection of a sensor measurement aimed at observing the state of a system which is described both by discrete variables and by continuous ones. The criterion is based on the expected value of the entropy variation associated to the sensor observation. This criterion is then applied to object recognition and localization tasks, in which the observed system is characterized by the object class, represented by a discrete variable, and by the object pose, i.e., position and orientation, represented by a vector of continuous parameters. The proposed criterion also accounts for the information obtained in the case the observed object is missed by the measurement.