Maixduino Powered Doorman

The Maixduino Powered Doorman is an edge-AI vision system designed to intelligently control automated doors using on-board machine learning. At the core of the setup is the Maixduino microcontroller, equipped with an integrated camera module and an LCD screen. The system runs a YOLOv2 facial detection model (flashed using Kflash_gui and programmed via MaixPyIDE) to analyze real-time video feeds directly on the device without relying on cloud computing. When a person approaches the entrance and looks toward the camera, the Maixduino processes the visual data and detects a valid face. Upon successful detection, the microcontroller outputs a control signal from its digital pins. Because the microcontroller's GPIO pins cannot directly drive high-current inductive loads, this weak control signal is fed into two BC545 transistors configured as electronic switches. These transistors amplify the signal to drive two 5V relays. The relays act as electrically isolated switches that can interface with external door-opening mechanisms. A dedicated 5V, 3A power supply provides stable, continuous power to both the Maixduino board and the relay switching circuit, ensuring reliable operation during active processing and relay actuation. By requiring a deliberate facial orientation rather than simple motion detection, the system prevents false triggers from casual hallway traffic.