The Adaptive Response of the Human Controller to Sudden Changes in Controlled Process Dynamics

A model is presented for the process by which a trained human controller detects, identifies, and adapts to changes in the gain and polarity of a compensatory velocity control system. The results of some experiments designed to test the model against actual human behavior are described. The features of the model which are explicitly tested are: 1) A detection process in which contemporary signal detection theory is applied to a variable which represents the difference between observed and expected system response to control movements; and 2) An identification process in which the human controller chooses that member of an ensemble of mental models of the controlled process which best accounts for the observed system behavior.