Three Models of Preview Control

This paper discusses a means to describe and eventually to predict the response of a human or artificially intelligent controller which 1) has a constrained preview of the actual input course and which 2) observes the successive target values as being of nonuniform importance. Three examples are: driving an automobile in traffic, a blind pedestrian using a cane or electronic obstacle detector, and remote manipulation of solid objects using artificial sensors and effectors. Three models are presented which characterize constrained preview control better than can conventional transfer function techniques.