2026 ASEE Annual Conference & Exposition

Abstract:
This paper describes the collaborative work of student teams in mechanical, electrical, and computing disciplines to design and build an automated solar-panel maintenance system. To sustain power production, the project tackles dust/debris buildup on panels. The prototype couples a wheeled cleaner, an imaging drone, and an AI pipeline that flags and follows contamination. The AI pipeline, a crucial component of the system, uses machine learning algorithms to identify and track the location of dust and debris on the solar panels. The effort is supported by a $40k grant and is scheduled for completion during the 2025–26 academic year; funding sources will be acknowledged in the final paper.
The Electrical and Mechanical Engineering Technology and Computer Information Systems teams, each consisting of an advisor of their majors and three students. The Electrical students who were deployed to work on the project began as paid student research assistants in Summer 2025 and have continued their work as part of two senior design courses. Educationally, the platform is being used in senior design and lab settings to reinforce core ECE/MET outcomes: power systems, motor control, embedded communication, systems integration, documentation, and test planning. In addition to the Electrical advisor, the instructor of two sequential Senior Design courses supports their collaboration by providing guidance and feedback. The Electrical Engineering lab assistant also assists with practical aspects of the project, highlighting the inclusive and collaborative nature of the project that involves multiple academic communities.
The completed phase spans both electrical and mechanical work. On the electrical side, the effort centered on power distribution and protection, motor control, and the embedded communications needed for wireless operation. On the mechanical side, the team configured the chassis, built the mounting hardware, and integrated a 6-DOF robotic arm. Validation included bench tests of current drawing, thermal limits, and subsystem reliability. Trials on mock solar panels confirmed stable power delivery, dependable wireless control, and repeatable arm positioning, all consistent with capstone rubrics.
Following the system workflow, validation proceeded step by step—from PC GUI input through Wi-Fi to the Raspberry Pi, then command parsing with serial handoff to the Arduino, generation of PWM/DIR signals, H-bridge drive of the DC motors, and finally encoder feedback returned to the GUI for live telemetry. Course-embedded assessments, including presentation and demonstration, are being used to document student performance. These assessments, designed by the instructors, evaluate the students' understanding of the project, their ability to apply theoretical knowledge to practical tasks, and their communication skills in presenting their work.
Looking ahead, the remaining work scheduled through April 2026—outside the results reported here—aims to enable closed-loop cleaning. This work is not just about innovation, but about making a real impact. The project's potential to enable closed-loop cleaning, autonomous navigation, and mechanical refinement could revolutionize solar-panel maintenance. The platform and accompanying assessment materials provide a reproducible ECE/MET building that other programs can adopt for laboratory use or senior design, further amplifying its potential impact.
Planned submission: Full paper for presentation.