KSEF RDE: A Dynamic Systems Approach to Understanding Human Learning

Humans possess exceptional learning capabilities. For example, humans learn to ride skateboards, fly kites, and ski. No existing control technique can replicate a human's ability to learn to interact with a wide variety of uncertain dynamic systems. The objective of our work is to identify beneficial learning mechanisms that humans use, but are not present in automatic control methods. The research outlined in this proposal provides an essential foundation for this long-term goal. In this one-year effort, we propose a series of human experiments designed to answer fundamental questions on the strategies that humans use to control dynamic systems. Specifically, we apply a dynamic systems approach to answer the following questions: In the control of linear systems (1) Do humans use the internal model principle? (2) Do human use high-gain control? (3) Do humans use feedforward control for command following? (4) Do humans use feedforward plant model inversion for command following? (5) What are the effects of system order on a human's ability to control a system? and (6) What are the effects of relative degree on a human's ability to control a system?