We believe there is a strong need in controls instruction and research for experiments as opposed to ``paper applications'' or simulations. Without experiments, students never see the gap between the mathematics and the engineering systems and they are never confronted by the leap of faith that engineers must make in any application. Without experiments, we do not learn the applicability of the theory, but more importantly, we fail to appreciate the limits of the theory, which is critical for both the theoretician and the practicing engineer. For this reason, well-designed experiments should be an essential part of control research and education. While this notion is now widely accepted, there remains in academia a significant split between theory and experimental application. This problem has many causes, but the principle one is that it is simply too difficult for theoretically-oriented faculty to do experiments relative to the perceived benefits in terms of tenure, promotion, and prestige.
A major objective of our project is to eliminate or reduce the obstacles to doing experiments within theoretical research and educational programs. Experiments alone will not replace the need to make contact with real engineering problems, but they can greatly facilitate such contact. Fortunately, an experiment can be made realistic enough by capturing the essential control issues without many of the complicating details of real engineering systems. The experiments that we plan on making a part of our curriculum at Caltech are good examples of this.