Abstract:

First-year electrical engineering students frequently struggle with validating basic circuits; misplaced components and wiring errors often go undetected, while inconsistent measurement interpretation hinders learning progress. Current MEXLE tools support structured experimentation but require manual verification. This paper presents two modules forming the core of a real-time circuit verification workflow:
(1) a computer-vision pipeline for automated circuit recognition on the MEXLEfirst board and
(2) a smart multimeter module for integrated voltage and current measurement.

The computer-vision pipeline automatically extracts circuit information from images of the MEXLE board. The pipeline performs ArUco-based geometric alignment, structured ROI extraction, YOLOv11-OBB/YOLOv11-based component detection, orientation estimation to identify flipped components, and OCR for reading component values. The output is a machine-readable description of component placement and wiring connections, enabling automated comparison against reference configurations or transformation into simulation-ready netlists. Experiments on a custom dataset of approximately 3000 ROIs demonstrate reliable detection across resistors, capacitors, wires, and diodes. Detailed accuracy metrics will be presented in the paper.

To complement the vision-based analysis, a compact multimeter module is developed for electrical validation. Designed for the MEXLE board’s shared I²C bus, the module measures voltages in the ±12 V range using high-precision ADC techniques and currents up to 2 A via an INA-based differential shunt stage. In addition, it includes electrical protection features against common measurement mistakes and can be flexibly placed at different board locations. A 32-bit microcontroller manages data acquisition and I²C communication, enabling integration with the vision pipeline and future MEXLEfirst workflows.

Together, these modules enable real-time automated circuit verification, reducing instructor workload and providing students with immediate, objective feedback. This paper focuses on the technical design and validation of the vision and measurement modules. Future work will combine the vision output with automated comparison against reference circuits, netlist generation, and multimeter-based electrical verification. Additionally, we will evaluate the impact on student learning and lab workflows forming a unified real-time assessment tool within the MEXLEfirst ecosystem.

Keywords:

Computer vision, automated circuit verification, electrical engineering education, embedded measurement system, technology-enhanced learning.