When engineers design systems like bridges or machines, they must ensure that the system and parts are strong enough to withstand the loads acting on it. Engineers often use a "free-body diagram" (FBD) to help with this design and analysis process. A FBD is a sketch of the system and the loads acting on it. It is important to teach engineering students how to draw and analyze a FBD correctly, but this is a difficult concept, can require a lot of practice, and often requires one-on-one feedback or grading from instructors. Before this research project, the only way to grade a FBD was manually by an instructor or other highly trained individual. Computers have not been able to perform this complex task because FBDs involve sketching, making correct assumptions about objects and the loads they experience, writing complex equations, and solving the equations. The time and effort required to provide one-on-one feedback is challenging for instructors, especially for large classes with hundreds of students.

Researchers on this project had previously developed software called Mechanix and shown how, using Artificial Intelligence (AI), it could recognize and understand a student's sketch, determine if it accurately represented the system, and if the student is evaluating the system correctly. This prior research showed that Mechanix had the potential to help students learn faster and more effectively by instantly giving them customized feedback. The goal of this new research project was to develop the additional complex algorithms and code necessary to accurately recognize and evaluate additional types of FBDs, to build Mechanix software into a complete online homework system, and to determine the impact of Mechanix on student learning. Researchers at Texas A&M University, Georgia Institute of Technology, Texas State University, San Jose State University, and LeTourneau University worked together to accomplish these goals.

The AI algorithms were successfully developed, tested, and built into a fully updated and rewritten version of Mechanix software. Instructors using the online platform can create homework assignments and view students' performance. Students can complete assignments by sketching a FBD, drawing appropriate forces, and writing equations necessary to evaluate the FBD. Mechanix is able to check student work for accuracy at various points in the process and can provide specific feedback instantly. Throughout this research project, trials of Mechanix were conducted in classes at all five universities with more than 1,800 students involved, and data on its effectiveness was carefully collected. Analysis of the data shows that the online Mechanix homework system is as effective as traditional, tried-and-true methods of evaluating FBDs and provides the feedback significantly faster. Because of the quick feedback, Mechanix also supports exploratory learning through creative engineering design problems where students can design bridge trusses to meet specific standards. These findings helped advance the fields of computer science and engineering education and resulted in many scholarly research papers and presentations in these fields.

LeTourneau University's role in this research project was using Mechanix in multiple engineering classes and collecting student performance data, such as homework scores and test scores. The data collected from students using Mechanix was compared to students who used a traditional homework system that included paper-and-pencil assignments and manual grading to understand the impact of Mechanix on student learning. Data from LeTourneau University was also compared to data from other institutions to understand if the impact was consistent at different universities, in different courses, and for different types of students. LeTourneau University also developed a bank of more than one hundred homework problems and solutions that could be used by instructors and students in this homework system.

Overall, this research shows that Mechanix is a helpful tool in the classroom to help reduce the burden of grading while providing feedback on student free body diagrams. Instructors using this system are able to provide students with feedback based upon scaffolded learning theories and have their homework assignments automatically graded for approximately the same amount of effort needed to set up a homework assignment in any online homework system. Further, Mechanix allows instructors to assign and grade open-ended problems, which would otherwise require a significant amount of grading time due to the possibility of each submission being unique. Students and instructors both seem to enjoy using Mechanix and students perform at least as well on Mechanix as other homework systems with the added requirement of drawing their free body diagrams.

Last Modified: 12/19/2023
Modified by: Benjamin W Caldwell