2026 ASEE Annual Conference & Exposition

This paper presents the implementation of the learning-by-doing pedagogical method into the design and development of smart systems integrated with open-source artificial intelligence (AI) tools for their operations and automated controls. The goal is to provide students with opportunities to be creative in researching and developing open-ended solutions to practical problems, while satisfying the requirements and constraints. Success observed, challenges faced, and lessons learned from case studies of capstone design and undergraduate research projects are discussed.
The first case is a capstone design project completed by two students majoring in Mechanical Engineering Technology (MET) and Electronics and Computer Engineering Technology (ECET). A remote-control vehicle equipped with a track system, a camera, and algorithms based on YOLO, and controlled by a Raspberry Pi single-board computer (SBC), was developed for surveillance and detection of targeted objects in unstructured terrains. Two teams, each consisting of three students majoring in Engineering Design Technology (EDT), MET, and ECET, are currently working on an industry-sponsored project to develop an automated non-lethal projectile launcher. The system automatically detects the existence and distance to targeted objects and then adjusts a launcher along both horizontal and vertical directions before launching the projectile. Currently, the students are conducting a literature survey to detect the controller, actuator for adjustments, and image analysis tools for object detection. The goal is to complete the system design by the end of the Fall 2025 semester and build a functional prototype by the end of the Spring 2026 semester.
With the support of the US Department of Agriculture (USDA), 15 undergraduate students majoring in MET, EDT, ECET, and Computer Science (CS) worked on a project to develop a smart device that can be used for on-site tests to determine whether a plant is infected with parasitic nematodes during the 2020~2024 academic years. TensorFlow and OpenCV are integrated into a Raspberry PI SBC, which controls a camera to take images of plant root systems, to detect the existence and species of parasitic nematodes through machine learning (ML), and to learn and identify infected plant root systems. Similar approaches have also been adopted in ongoing research to equip a three-degree-of-freedom (3DOF) DexArm robotic arm by Rotrics with computer vision for automated material handling. The autonomous mobile robot LIMO, developed by Agilex Robotics and controlled by an Nvidia Orin Nano running ROS, is also being developed to have environmental sensing capabilities. Computer vision is to be developed for the 6DOF collaborative MyCobot 280 robotic arm, mounted on the LIMO, to create a compound platform with flexibility for various automated functions. The goal is to integrate these smart robots with other manufacturing equipment for a fully automated manufacturing system.