Using Project Based Learning (PBL) with Control Theory

Abstract

This paper is an Evidence Based Practice submission for Project-Based Learning of Automated Control Theory. The approach uses PBL rather than a traditional lecture and exam to evaluate student learning. Control theory and system identification can be very mathematically heavy. Whether analyzing these systems with Laplace transfers, in State Space or with differential equations in the time domain these problems can be tedious. Not surprisingly, it is often difficult for students to grasp the direct effects of the mathematical methods on controlling the system or plant. In other words, it can be difficult for students to fully grasp how to physically change their system in the same manner that mathematics does. To mediate this source of confusion most control course curriculums include some form of laboratory experiments. More often than not this requires the learning of new software like LabView to drive the control experiments. It also includes other hardware that can be equally expensive. This can leave students dependent on expensive hardware and software to control even the simplest of systems. In this paper we compare two approaches; one with MATLAB and another more direct approach with an Arduino based controller. In both cases the students need either prior knowledge of MATLAB or Arduino programming to complete the experiments. For the see-saw experimental apparatus that we developed, we provide the details and costs associated with making and developing these experiments. We show the contrast between the two experiments using student surveys and provide the results. We also examine the differences between student comprehension as a measure of merit of the two methods. In particular we examine a low-cost 1st order temperature apparatus developed at BYU and then develop a 2nd order see-saw balancer of our own. In both cases these devices cost less than $50.00 to produce and our students were able to build these without much difficulty. The usefulness of physical models in control curriculums cannot be underestimated. Students getting hands-on experience controlling mechanisms or circuitry benefit from these real-world experiences. We provide the data taken from the students as evidence of merit.