Stimulating Higher Order Thinking in Mechatronics by Comparing PID and Fuzzy Control

TL;DR

The paper addresses the challenge of cultivating higher-order thinking ($analysis$, $synthesis$, $evaluation$) in engineering education by designing a semester-long mechatronics course that culminates in an open-ended comparative study of PID and fuzzy control for a leader-following robot. The approach combines hands-on lab work (Arduino-based controllers, PID and fuzzy implementations), computer vision (Pixy), and real-world sensing with data-driven evaluation, guided by in-progress reviews and scholarly writing training. Key contributions include demonstrating that students can create their own evaluation criteria and experiments, and showing that open-ended comparisons are technically nontrivial due to multiple confounding factors. The work suggests a practical framework for fostering higher-order thinking in senior undergraduates through real-world, interdisciplinary projects and peer-driven critiques, with findings supported by multi-semester experience and student surveys.

Abstract

Many studies have found active learning, either in the form of in-class exercises or projects, to be superior to traditional lectures. However, these forms of hands-on learning do not always lead students to reach the higher order thinking skills associated with the highest levels of Bloom's Taxonomy (analysis, synthesis, and evaluation). Assignments that expect students to follow a prescribed approach to reach a well-defined solution contribute to a lack of higher order thinking at the college level. Professional engineers often face complex and ambiguous problems that require design decisions for which there is no straightforward answer. To strengthen the higher order thinking skills demanded by such problems, we developed a project in a semester-long mechatronics course in which students must evaluate two automatic control methodologies without being given explicit performance criteria or experimental procedures. Specifically, the project involves determining the superior control method for leader-follower behavior, where a ground vehicle autonomously follows a lead vehicle. Laboratory exercises throughout the semester expose students to the skills required for the project, including using sensors and actuators, programming proportional-integral-derivative (PID) and fuzzy controllers, and applying computer vision to detect an object signature. In the final course project, students go beyond implementing individual controllers and create their own evaluation criteria and experiments to make a design decision between PID and fuzzy control. We implemented this approach over three semesters and found that students value working on a real-world, open-ended problem, develop creative performance criteria and evaluation methods that demonstrate higher order thinking, and discover that comparative studies are nontrivial due to the many factors influencing performance.

Figures (10)

• Figure 1: The Mechatronics course is typically taken in the first semester of the senior year. It is also part of the Robotics minor. Almost all students that enroll are electrical engineering or mechanical engineering majors.
• Figure 2: Snapshot taken from the camera's software known as Pixymon. Pixymon can be used to configure the camera to recognize different colored objects and identify them uniquely. Once the camera is configured to recognize a particular color, it automatically bounds that color with a box like the one shown in the figure. The position and size of the box represents the pixel-level information that is sent to the microcontroller for control (tracking) decisions.
• Figure 3: Picture showing a pair of leader-follower robots in the hallway of our academic building. The camera mounted on the front of the follower vehicle is visible. The red paper on the leader is an easy-to-find target for computer vision.
• Figure 4: Illustration showing the typical experimental setup developed by the students to subject the speed controller to different types of step responses.
• Figure 5: Illustration showing one of the challenges encountered when some students attempted to evaluate their steering controllers. They discovered that the controller either did not have the opportunity to correct its alignment or the follower ended up behind the leader, but at an angle that would eventually need counter correction.